Why "that would be me"? (part 1)

« previous post | next post »

There's a recently-fashionable construction, in which "would be" is used where plain "is" might have been expected. For example, in the imaginary Q&A below, I might respond with B2 rather than B1

A: I'm looking for Mark Liberman.
B1: That's me.
B2: That would be me.

A couple of weeks ago, our comments section featured a lively discussion of this phenomenon. (As far as I know, there isn't any common-used term for it, so pending a better idea, I'll call it the TWBM construction, for "That Would Be Me"). Opinions differed, as they often do in discussions of matters linguistic, about where to draw the boundaries of the phenomenon, as well as about its meaning, origins, circumstances of use, and so on. In particular, Bloix suggested that "The point of the 'would be' construction is that it implies doubt on the part of the speaker", while I expressed skepticism about the relevance of doubt to the meaning of this construction.

A couple of small recent experiments have tested this question. Their results suggest that effect of would on perceived speaker certainty in TWBM sentences is very small — too small to be a sensible motivation for using this construction. And also, the small effect, to the extent that it could be determined to exist at all, was positive. That is, in one experiment, the effect was not statistically significant; and in a second experiment, speakers using would be were judged, on average, to be expressing a bit less doubt than those using is.

The research in question was mentioned (though not described) in a guest post by Larry Horn that I linked to in my post of a couple of weeks ago. Larry's post was "That would be in the modal auxiliary, Bob", 1/9/2008), and it referenced a 2008 LSA talk entitled "The effect of semantic modality on the assessment of speaker certainty". That LSA talk doesn't seem to exist in any more permanent form, but the same authors offer what appears to be similar material as Augustin Gravano et al., "The Effect of Contour Type and Epistemic Modality on the Assessment of Speaker Certainty", Proceedings of 4th Speech Prosody Conference, Campinas, Brazil, 2008. Their abstract:

In an empirical study participants were asked to rate the perceived degree of certainty of utterances that contained either the modal would or main verb be (e.g. That would be me vs. That’s me), and which were also variously produced with one of three intonational contours (downstepped, declarative, and yes-no question). We found that both downstepped contour and epistemic would made a significant and independent contribution to the assessment of speaker certainty. That is, participants rated utterances with the downstepped contour as most certain, followed by those with the declarative contour, while the yes-no question contour was perceived as highly uncertain. Similarly, participants rated speakers' responses with epistemic would as significantly more certain than those without it.

The paper actually describes two experiments. In the first one, 12 NWU undergraduates judged the certainty of would be and is in written scenarios, using a 5-point Likert scale (-2 = very uncertain, -1 = somewhat uncertain, 0 = neither certain nor uncertain, 1 = somewhat certain, 2 = very certain). Their judgments were transformed to z-scores:

zij = (Xij-mean(Xi))/stdev(Xi) ,
where Xij is the jth judgment by subject i,
and Xi is the set of all judgments by subject i.

In this first experiment, the paper reports that

The mean Certainty for stimuli with modal would was -0.13 (st.dev. = 1.11), and for stimuli without modal would was -0.03 (st.dev. 1.04).

This average difference of a tenth of a standard deviation, in a small experiment, was not statistically significant. More important, the small size of the average difference, and the large overlap in judgments, means that this experiment casts doubt on the idea that the dimension of speaker (or writer) certainty is relevant at all in choosing to use the TWBM construction.

In the second experiment, the same scenarios were spoken rather than written, and the TWBM-type punch lines were produced with three different intonations. A somewhat larger set of undergraduates performed as subjects. In this experiment, the average effect of would be on certainty judgments was statistically significant — and positive — though again not large enough to be practically significant:

(The "Certainty" scale is in terms of z-scores again.)

There are several problems with experiments of this type. One problem is that a small set of students, typically from a single class, is taken as representative of the entire English-language speech community. Especially in matters like the TWBM construction, where there have apparently been recent changes in usage, it would not be surprising to find regional, generational, class, or ethnic differences. A second problem is that focusing on a single dimension of variation can often create artefactual patterns of judgment — participants need to create and apply, on the spot, a theory about the particular set of variants that they are hearing over and over again, in the particular set of contexts in which they're hearing them, so that the experiment may be testing their category-learning propensities rather than their pre-existing linguistic knowledge. And finally, conscious judgments about when a form would be used, or what it would mean to use it, are often strikingly different from the patterns actually observed in the linguistic behavior of the individuals making the judgments.

For all these reasons, we shouldn't take such experiments to be definitive. Still, I'm going to take this one as tending to support my view that expressing "doubt", ironic or otherwise, is not what the TWBM construction is all about.

But if the TWBM construction is not expressing doubt, what *is* it doing?  In order to address that question, I'll start with the description of would in the Cambridge Grammar of the English Language (which doesn't include any specific discussion of TWBM examples). I'll continue with a small corpus study of would, looking at the distribution of the CGEL categories, and noting the existence of a common usage that seems to have been left out, or at least under-emphasized. And I'll end by observing that the TWBM construction can be seen as a special case of this common use of would.

I won't do any of that today, however — tune in again tomorrow.


  1. Ray Girvan said,

    April 2, 2009 @ 7:11 am

    I agree it seems part of a wider pattern of using "would" (such as in giving advice or answering questions):

    A: I found a corpse in my dustbin yesterday; what should I do?
    B: I would say, go to the police.

  2. Marc Hamann said,

    April 2, 2009 @ 7:38 am

    The first thing that comes to mind about this construction is that it has the advantage of "prescriptivist avoidance".

    Anyone who learned in school that you were supposed to say things like "It is I", but who don't really believe that sounds right can avoid the whole problem by using the modal.

  3. Vincent said,

    April 2, 2009 @ 7:43 am

    From an especially British perspective, I suspect that the construction expresses a sarcasm or false modesty, designed as a figleaf to conceal the speaker's actual glee at knowing the correct answer. Thus when correcting someone's mistake, an Englishman might start with "I think you'll find that …"

    It strikes me as somewhat analogous to challenging a trespasser (even perhaps a burglar) with the words "Can I help you?" delivered in a certain tone of voice: "haughty self-assurance" in John Caldwell's phrase.

    It might be translated as "I am of course too well-bred to be triumphalist or aggressive, but that doesn't stop me achieving my desired effect, with a lift of the eyebrow, a tone of voice or a well-honed phrase."

  4. Ray Girvan said,

    April 2, 2009 @ 8:03 am

    From an especially British perspective, I suspect that the construction expresses
    I agree, under some circumstances. But (again UK English, early 50s speaker) to me it's a neutral frame for expressing an opinion, and often contracted to very rapid "I'd say".

    Mrs G: What do you want to do with the X that Y sent us?
    Me: I'd say keep it / give it to the charity. *

    * dual options to show no positive/negative weight attached.

  5. Tim Silverman said,

    April 2, 2009 @ 8:20 am


    I think we probably need to distinguish a "straight" from an "ironic" usage.

    E.g. Straight–

    Visitor: I need to speak with your local butterfly expert.
    Native: Hmm, I suppose that would be Dr Sniggleton.

    vs Ironic–

    First flatmate: Where's the tea kettle?
    Second flatmate: That would be that tea kettle-shaped object on the counter in front of you.

    The ironic version has become popular, but presumably it gets its force from the straight version. I do think there's something to do with uncertainty in the straight version, since the thing that would stop me from saying "That's Dr Sniggleton" rather that "That would be Dr Sniggleton" is basically my uncertainty that that is the right answer, or at least a feeling that the right answer is not obvious. Then again, it might be used for the sake of politeness; perhaps to avoid showing up the ignorance of the questioner, by pretending that the answer is not obvious to the speaker either (even when the ignorance of one and the knowledge of the other may be perfectly reasonable in the circumstances—politeness can be like that).

    The ironic usage would then play off the fact that the answer is perfectly obvious, and hence the pretence that it is not gives a comic effect.

    There is also plenty of room for a grey area in between, where the level of obviousness is not clear, perhaps as the ironic usage spreads and weakens.

  6. Karen said,

    April 2, 2009 @ 8:26 am

    There's also a pragmatic politeness effect. "That would be me" – when spoken straight, ie, not when I'm wearing a huge name tag – slightly distances me from an implication that you should have known it. I think, in fact, that's why the ironic one works: often, you *should* have known it, or could at least have reasonably been expected to guess. Not always – a visitor encountering several people in the lobby could probably be expected NOT to know which is me.

  7. Kenny V said,

    April 2, 2009 @ 8:31 am

    In the TWBM construction, my gut tells me that it has something to do with not wanting to use "is" with the first person–so you take "would," which doesn't conjugate.

  8. Stephen Jones said,

    April 2, 2009 @ 8:40 am

    Couldn't it be a simple example of social distancing?

  9. Adam Roberts said,

    April 2, 2009 @ 8:46 am

    I have a different sense of this. 'Would' might be being used here as a form of rhetorical unclarity, which in turns works to connect with the questioner, a sort of 'I'm with you' device. 'I'm looking for Mark Liberman' implies 'I don't know Mark Liberman or where he is' which in turns announces the speaker's ignorance and uncertainty. Saying 'that's me' trumps the speaker's uncertainty with certainty, and there are situations where doing that too assertively would border on rude. Saying 'That would be …' moves the statement back onto the border of uncertainty (even though Mark Liberman of course knows who Mark Liberman is); because '…would be…' is less deictic and assertive than '…is…' In other words, I suggesting this formulation is parallel to, eg:

    A: I'm looking for Mark Liberman.
    L: I guess that's me.

    This isn't a million miles from what Karen says, above.

  10. Kris Rhodes said,

    April 2, 2009 @ 9:25 am

    My guess–and it's only a guess–is that "that would be me" expresses a tentativeness, not regarding the identification, but regarding the _consequences_ of the identification. In other words, it means roughly "that would be me [but let's check (or I'm willing to check) to make sure that identification has the significance you think it does]."

  11. Phil Goldfarb said,

    April 2, 2009 @ 9:55 am

    My impression, at least when I think about it, is that I, and most of my acquaintances, use "would be," when not referring to ourselves, when the answer requires some thought or recollection. So "Who is the best basketball player here?" "That would be John," with the "would be" serving as a placeholder for "gimme a second…oh, here's the answer." So I would suggest that the "would be" in the TWBM construction serves as politeness (or sarcasm) by suggesting that the matter requires at least a little thought because it is non-obvious. I suppose that's fairly similar to Karen and Adam Roberts' suggestion – it's pushing away the certainty for some rhetorical effect.

  12. Nada said,

    April 2, 2009 @ 10:16 am

    Many questions are answered with "would", especially semi-rhetorical ones that are just looking for confirmation, and also, statements of uncertainty looking for confirmation.
    ("I think I'm going to X."/"That would be the best thing to do.") Maybe "that would be me" is a particularly common form of this asserting "would", but I wouldn't have thought so before this post.

    I do have a sense that "would be" is used when uncertainty is floating about, like the other commenters say. It comes out a lot in a tech or mechanical environment – "I'm hearing this clankity-clankity-clank every time I turn a sharp right," "Hmm, that would be your carburetor dragging on the ground," or "My computer suddenly turns itself off in the afternoons," "I think that would be your fan – it sounds like you're overheating." There, I think "would be" happens out of cautious skepticism – the tech may not actually have all the information. The statement is secretly, "If this were all the data about the situation, then the answer would be…" So, I wonder if "that would be me", specifically, comes out of situations where it's uncertain whether I am actually the person being sought. "I'm looking for the person in charge here!", "I'm looking for Jimmy!", etc. The answer is "If I am the person you want, then [the identifier you gave] would be me," essentially.

    That's a possible origin story – then it becomes idiom and people use it with no special intent.

    I think that as an extension of the "confirmation" cases above, "that would be x" is also used sometimes to be smarmy:
    "What's this doohickey?"
    "That would be the 'on' switch."
    … but I think smarminess is an extension made possible by the main uses.

  13. marie-lucie said,

    April 2, 2009 @ 10:25 am

    "That would be me" (implying: "if I am indeed the right person") would have been a good reply in a situation I encountered some time ago. I was sitting in my office, the door of which displays my full name, when a student came in looking for (my name). "That's me", I said. He looked confused and asked if there was another teacher by that name. "No, I am the only one. What seems to be the problem?" It turned out that this student, registered in my section but having never shown up in class, had been (sporadically) attending a section of a different course, taught in the same time slot but by a teacher whose first name had the same initial as my last name (neither of those names being at all common). People having relatively common names might well reply "That would be me" in the knowledge that they might have the right name but not be the person actually sought.

  14. Arnold Zwicky said,

    April 2, 2009 @ 10:31 am

    Marc Hamann: "Anyone who learned in school that you were supposed to say things like "It is I", but who don't really believe that sounds right can avoid the whole problem by using the modal."

    Not really; there's still a choice between "I" and "me".

    In fact I doubt that even severe prescriptivists would insist on "That's I" (rather than "That's me") in identifications; in my experience, prescriptivists are unwilling to carry an insistence on nominative predicatives very far from a few formulas like "It is I".

    But someone who insists on "That's I' would presumably also insist on "That would be I".

  15. Dierk said,

    April 2, 2009 @ 10:41 am

    I am with the current minority: It's politeness.

    Often enough we [linguists, scientists] forget that language is about communication, which in turn is a social interaction. Part of this is a game in which the participants try to figure out the social status of others and of themselves in a certain context.

    Politeness is a fall-back position when one has no indication of how friendly, smarmy, arrogant, snotty, hostile etc. one can be without reprimands. 'That's me' constructions are highly declarative, and can sound mildly to very brash – especially in a society in which certainty has been lost to a great degree.

    At the same time people want what they think is 'factual language', that is, short sentences, easy to understand without any ado 'coming to the point'. Typical for advertising, which has had a large influence over the past 30 years. The long polite version is thus out, most people will be bored and taken aback if you come up with 'If you are looking …' leading to the 'That would be me'. Polite but short.

  16. Russell said,

    April 2, 2009 @ 10:45 am

    I haven't read any of the posts in question (sorry), but I didn't see mention of this article, so I'll mention it. Apologies if it's old news.

    Birner, Betty, Jeffrey P. Kaplan, and Gregory Ward (2007) Functional Compositionality and the Interaction of Discourse Constraints. Language, Volume 83, Number 2, June 2007, pp. 317-343.

    In which they syntactically and pragmatically decompose "that would be X" into its component constructions (including epistemic "would")

  17. Ian Tindale said,

    April 2, 2009 @ 10:46 am

    I consider it a way of pointing out that the other person has asked a slightly stupid question or is a slightly stupid person asking a question, and that if they were in any way up to normal standards of intelligence they'd have found out the answer that we're giving them themselves, as it's so utterly obvious as to be staring them in the face (even if mostly it isn't).

    Jo Brand is a master of this sort of tool in her humour toolbag. Or at least, she was about 15 years ago. Now, just about everybody who works in an office in central London seems to be a master of this sort of tool in their humour toolbag.

    "Which train goes to the ExCeL centre?" asks a visiting G20 world leader.
    "That'd be the Beckton train", I'd say, looking directly at the DLR map in front of us.

    I suppose you could substitute the "that would be" back to "that is" and go one simple step further and simply put "duh" at the end of the sentence. It amounts to more or less the same thing, only stronger, stupid.

  18. Fluxor said,

    April 2, 2009 @ 10:50 am

    In the example given at the beginning of this post, it appears to me that there is an implicit question being asked.

    A: I'm looking for Mark Liberman.

    A1: I'm looking for Mark Liberman (and is that you, perhaps?).
    B1: That's me.

    A2: I'm looking for Mark Liberman (and would that be you, perhaps?).
    B2: That would be me.

  19. K said,

    April 2, 2009 @ 11:23 am

    I agree with Karen, that it seems somewhat polite. It implies that the person seeking the information has some doubt and it's as if, by using would, you are acknowledging their doubt and getting onto the same level as them. It may also come from the widely used question of "You wouldn't happen to be X, would you?"

    I agree with fluxor about the implied question being asked, and I think I would use the answer of "That's me" and "That would be me" depending on the level of doubt from the speaker. If a person says to me "I'm looking for Dr. Green" and it seems as if there is no way that he/she assumes that is me, I would say "That's me." But if someone came in saying "I'm looking for Dr. Green" and it seems as if they think there is a chance he/she believes it is me (and therefore reminding me of a construction with pragmatic meaning similar to that of "You wouldn't happen to be Dr. Green, would you?" then I would be more likely to respond with "That would be me".

  20. Harry said,

    April 2, 2009 @ 11:24 am

    The most strident example I've noticed is "I would argue that…" used to open an argument that someone is actually making.

  21. Josh said,

    April 2, 2009 @ 11:24 am

    The "doubt" expressed by the TWBM isn't uncertainty about the answer to the question, so much as uncertainty (polite, feigned, or otherwise) about the question itself. In your original example, the first speaker doesn't even ask a question. Therefore, the second speaker replies by giving an answer that is conditional upon the question that he assumes to be implied:

    A: I'm looking for Mark Liberman.
    B: [If you're asking who Mark Liberman is,] that would be me.

    The implication of an unasked question is also why the TWBM construction is so useful for irony:

    A: What's this doohickey?
    B: [If you're really so stupid you have to ask,] that would be the 'on' switch.

    The role of the TWBM in unironic answers to direct questions is a little less clear, but can still be seen as an implied hypothetical used to introduce some rhetorical distance to show modesty or genuine uncertainty:

    A: What do you think about this movie?
    B: [If you really want my opinion,] I'd say it's okay. OR [If I have to have an opinion about it,] I'd say it's okay.

  22. Robert Daland said,

    April 2, 2009 @ 11:24 am

    I will echo and amplify Russell's comment:
    Gregory Ward has worked extensively on this construction. His homepage is http://www.gregoryward.org/publications.html

    Here are some publications that deal with this and related constructions:

    2008. Ward, Gregory. “Equatives and Deferred Reference,” in Reference: Interdisciplinary Perspectives, edited by Jeanette Gundel and Nancy Hedberg. New Directions in Cognitive Science Series. Oxford: Oxford University Press. Pp. 73-92.

    2007. Ward, Gregory, Jeffrey P. Kaplan, and Betty J. Birner. "Epistemic would, Open Propositions, and Truncated Clefts," in Topics on the Grammar-Pragmatics Interface, edited by Nancy Hedberg and Ron Zacharski. Amsterdam/Philadelphia: John Benjamins. Pp. 77-90.

    2004. Ward, Gregory. “Equatives and Deferred Reference,” in Language 80:262-289 (http://www.ling.northwestern.edu/%7Eward/Language04.pdf)

    2003. Ward, Gregory, Betty J. Birner, and Jeffrey P. Kaplan. “A Pragmatic Analysis of the Epistemic Would Construction in English,” in Modality in Contemporary English, edited by Roberta Facchinetti, Manfred Krug and Frank Palmer. [Topics in English Linguistics 44, General Editors: Bernd Kortmann and Elizabeth Closs Traugott.] Berlin/New York: Mouton de Gruyter. Pp. 71-79. (http://www.ling.northwestern.edu/%7Eward/MCE03.pdf)

  23. Ransom said,

    April 2, 2009 @ 11:26 am

    The whole thing reminds me of future tense in German. From "German in Review" 3 ed. (Sparks, Vail) 2.II (p. 44):

    "Conversational German tends to use the present tense to express future time. However, German does use the future tense in the following cases:

    "3. Werden can be used to indicate an assumption about something connected with the present time. This will occur in such a context as:
    'Wer klopft an die Tuer? — Das wird Peter sein.'('Who is knocking at the door? — That'll be Peter.')"

    The parallel is perhaps not exact, but I find that the construction has a similar feeling for me.

  24. marie-lucie said,

    April 2, 2009 @ 11:46 am

    Several of the comments here seem to consider that an answer including would implies that the question prompting the answer is considered stupid. Simple, legitimate ignorance or unfamiliarity, as in the case of the hypothetical G20 attendee or other stranger to the area, is not even a possibility. So I suppose such a question can only be considered legitimate and receive a straight, non-would answer if it is asked by a young child.

    On the other hand, the other condition or presupposition to the answer "that would be X" could be "If I am not mistaken", rather than "If you are so stupid as to ask" – false modesty sometimes perhaps, but also possibly genuine acknowledgement of one's fallibility.

  25. bianca steele said,

    April 2, 2009 @ 12:44 pm

    I agree with Karen and, I think, Adam. "Would" is deferential to the questioner and declines to assert too strongly anything he does not seem to know, implying that any moron could infer the conclusion from the knowledge he already has, even where this is obviously not the case.

    Although Ian implies that in British usage, the sarcasm trumps the politeness: why should one distance oneself from an implication that you should be able to get the information yourself without asking.

  26. Kim said,

    April 2, 2009 @ 1:48 pm

    The context I've heard TWBM in is similar to what Marie-Lucie described. The asker is looking for someone they don't know. The answer is sometimes "That would be him/her", and sometimes "That would be me".

  27. Ray Girvan said,

    April 2, 2009 @ 2:52 pm

    Arnold Zwicky: in my experience, prescriptivists are unwilling to carry an insistence on nominative predicatives very far from a few formulas like "It is I".

    Sorry to say, see here on Yahoo Answers where one Mistah Rickstah (who is generally worth following for anyone who wants to boggle at a spectacle of bizarre and arbitrary prescriptivism) seriously argues that "Maybe it's just me" should be corrected to "Maybe it's just I".

  28. Rubrick said,

    April 2, 2009 @ 4:06 pm

    I think "that would be me" introduces a tone of mock-pomposity or pedagoguery. It feels very similar to the use of the phrase "this is true" to indicate agreement, as opposed to "that's true" (a usage which feels like a last-couple-of-decades phenomenon to me, at least in the U.S., but that might well be the Recency Illusion).

    However, I'd hazard that the main reason people say "that would be me" or "this is true" is that they've heard other people use it and it feels kind of fun. People just enjoy changing things up now and again.

  29. Faldone said,

    April 2, 2009 @ 4:46 pm

    @Ray Girvan

    One could always argue back ot the Mista Ricksta types that, at least in the "is it just me" construction the "me" is dative, something akin to the "woe is me" construction.

  30. Nathan Myers said,

    April 2, 2009 @ 5:06 pm

    I would never accuse a conflagration of linguists of over-analyzing a grammatical construction. Still, people who use the construction, if you ask them, are not usually imagining snarky unspoken subclauses. Doesn't it suffice to say that changing "is" to "would be" adds distance? Often adding distance is trivially polite, other times it reduces commitment, leaving room to backpedal. These motivations seem to me extralingual, all manifesting more or less indistinguishably as the thoroughly linguistic phenomenon of distancing.

  31. Fred said,

    April 2, 2009 @ 11:46 pm

    I think Nada is onto something. A similar distinction is made in German, no?

  32. Bob Kennedy said,

    April 3, 2009 @ 12:23 am

    I think there's a speech perception angle to be explored here as well. The utterance "That would be me" is longer than "That's me" or just "Me", and this length could increase the likelihood of the response being understood as contentful rather than some other vocalization like a cough or hiccup. Plus it could help the addressee calibrate the vowels of the speaker. Not that these are the only reasons someone would use this phrase, but it's a plausible additional motivation.

  33. ChrisB said,

    April 3, 2009 @ 12:53 am

    And what a splendid opportunity to put in one of the few Dorothy Parker jokes where she comes out one down:
    The newcomer being shown round the New Yorker who had Robert Benchley pointed out to him, then indicated the women with him and asked, "Would that be Mrs. Benchley?"
    "Yes, it would… but it is Dorothy Parker."

  34. Fred said,

    April 3, 2009 @ 1:09 am

    Oops, hit Submit too soon… in particular, this reminds me of Indirekt Rede in German subjunctive:


  35. dr pepper said,

    April 3, 2009 @ 2:52 am

    Remember a few years back when there was a fad for commercials featuring dead celebrities?

    In one, a man gets into what looks like the beginning of an unfortunate confrontation with two other men in a bar. When he says something assertive, one of the others says, condescendingly, "who do you think you are, John Wayne". From behind the two troublemakers, a voice answers "no– that would be *me*". Seeing the real John Wayne, they sputter and quickly leave.

  36. Gregory Ward said,

    April 7, 2009 @ 11:06 am

    Thanks both to Russell and to Robert Daland for pointing readers to our work on what we (Betty Birner, Jeff Kaplan, and I) call ‘TWBX’ (Mark’s ‘TWBM’). As Robert noted, we have published extensively on this construction, including the 2007 article in Language article cited by Russell.

    Our central conclusion about the TWBX construction (and epistemic 'would' in general) is that its felicitous use requires a contextually-salient open proposition (OP), in the sense of Prince (1986). Epistemic 'would' is unique among the modals in this respect. In addition, the subject demonstrative that in TWBX can be used to refer deictically to the instantiation of the variable of the OP, allowing the speaker to equate it, via the equative syntax, with the post-copular focus. Furthermore, a use of TWBX marks the instantiation of the variable (in the post-copula focus) as exhaustive.

    Mark’s posting and the subsequent responses develop a number of themes, among which are:

    • whether TWBX marks certainty or uncertainty'
    • the ‘haughty self-assurance’ it frequently conveys;
    • whether it presupposes a question

    Our work actually began from the observation of the certainty conveyed by TWBX, in examples like some of those cited by Mark and responders, e.g., Vincent’s example:

    First flatmate: Where’s the tea kettle?
    Second flatmate: That would be that tea kettle-shaped object on the counter in front of you.

    This is consistent with the conclusions of Gravano et al. (2008), who found in their empirical study (also referenced by Mark) that participants judged utterances produced with epistemic 'would' to be more certain than the corresponding utterances without the modal.

    The ‘haughty self-assurance’ that is sometimes conveyed (as in Vincent’s example) we suppose derives from the high level of certainty conveyed. Another derivative attitude/stance frequently associated with the use of the construction is humor/sarcasm, as in:

    [A accidentally steps on B’s foot]
    B: Um, that would be my foot.

    Here the humor follows from the fact that B’s uses a construction associated with certainty to identify the object in question (his foot) about which there is presumably no uncertainty (especially in light of the fact that someone has just stepped on it).

    As for the presupposition of a (possibly only implicit) question or question-under-discussion (QUD), at one point we considered this approach but rejected it in favor of our (simpler and more general) OP analysis, especially since some uses of TWBX preclude the presupposition of a question, e.g., “correctives”:

    Villager (in reference to an ogre): He’ll grind your bones for his bread!
    Shrek: Actually, that would be a giant.

    Here, the Villager’s utterance does not make salient the question 'Who’ll grind your bones for his bread?' or anything similar; instead, it makes salient the open proposition 'Someone grinds your bones for his bread'.


    Birner, Betty J. Jeffrey P. Kaplan, and Gregory Ward. 2007. Functional compositionality and the interaction of discourse constraints. Language 83 (2).317-343.

    Gravano, Agustín, Stefan Benus, Julia Hirschberg, Elisa Sneed German, and Gregory Ward. 2008. The Effect of Prosody and Semantic Modality on the Assessment of Speaker Certainty. Fourth International Conference on Speech Prosody, Campinas, Brazil. May.

    Prince, Ellen F. 1986. On the syntactic marking of presupposed open propositions. In A. Farley, P. Farley, and K.-E. McCullough, eds., Papers from the Parasession on Pragmatics and Grammatical Theory, 22nd Meeting of the Chicago Linguistic Society. 208-22.

  37. Elisa Sneed German said,

    April 15, 2009 @ 4:35 am

    To piggyback on Gregory’s comment, I would add that while our findings may not be “definitive” (and what findings are?) and despite the limitations of this series of studies, we believe that experimental research in this domain is a worthwhile and valuable endeavor.

    There are limitations to any experimental research, not the least of which is the available population of participants. The participants in these studies were taking an entry-level linguistics course at Northwestern, a partial requirement of which was to participate in two linguistics experiments. University statistics for the class of 2012 indicate that 37% of students are from the Midwest, followed by 20% from the mid-Atlantic and 14% from the West. (Ten percent or less come from each of the following regions: the South, other countries (i.e. International students, excluded from our analyses), the Southwest and New England). The information the University tracks on ethnicity indicates that 21% are Asian-American, 7% Hispanic, 4% African-American and <1% American Indian. (Data on the ethnicity of the remaining 67% are not provided.) Most of our participants were female. So it is certainly possible that the data from our 42 participants reflect a regional, ethnic and gender bias in addition to the more obvious generational bias that results from using all undergraduates.

    As far as creating artefactual patterns of judgment is concerned, this too is certainly a risk; it is perhaps a greater risk for the written study, where subjects were asked to evaluate only the certainty of the TWBX construction without explicit prosody. However, in our oral experiment, where we found significant effects, participants were asked to evaluate our test items when both prosody and construction were varied. Further, we used twice as many filler items as test items in order to best minimize the possibility that any effects we obtained would be due to learning during the experimental session.

    Finally, asking participants to make conscious judgments about a linguistic stimulus is necessarily different from observing naturally generated linguistic behavior. The fact remains that when these participants were presented with the TWBX construction and either the downstepped or declarative intonation contour they regularly responded in a different way than they did when presented with indicative be and either of these contours (or when presented with the yes-no question contour). The fact that we got these results, then, provides evidence that — at least for the participants in our study — the interpretation of certainty is conventionally associated with the epistemic modal would.

    In the absence of other experimental research on this topic, the results of this series of studies provide a good working hypothesis about the meaning of the TWBX construction and its interaction with prosody.

  38. The Connoisseur said,

    November 23, 2010 @ 9:41 am

    The expression comes from such questions as “Who would be the person that I see to order tickets?” The answer can be “That (person) would be I”, in other words, “I would be that person”. Or the answer can be “that (person) would be me”, in other words, “You would see me.”

    This is similar to the “It is I/me” debate. In answer to the question “Who is it (that knocks)?”, the answer can be “It is I” or “I knock.” However, in answer to the question “Whom is it (that I hear)?”, the answer can be “It’s me (whom you hear)” or “You hear me.” Few people ask the question “Whom is it?”, but both answers are given to the question “Who is it?

RSS feed for comments on this post