Why "that would be me"? (part 2)

« previous post | next post »

As promised in part 1, I'm going to survey CGEL's taxonomy of uses of would, and do a tiny corpus study to get an idea of their relative frequency.  In a later post, I'll take up the implications for the recently-fashionable "that would be me" construction. (For background, see "We've met the enemy, and that would be in the modal auxiliary, Bob" 3/18/2009, and "Why 'that would be me'? (part 1)" 4/2/2009.)

The Cambridge Grammar of the English Language, Chapter 3, The Verb, by Rodney Huddleston, covers "The preterite forms could, might, would, should" in section 9.8, pp. 198-302. The section starts this way:

We have distinguished three uses of the preterite: past time, backshift, and modal remoteness. It is a distinctive property of the modal auxiliaries that the modal remoteness use is much more frequent and less restricted than the past time use — the complete reverse of what holds for other verbs.

What do these three uses mean in the case of the modal auxiliary would, according to CGEL?

1. Past time would. This use is typically associated with the senses of volition or propensity: "I had no money but he wouldn't lend me any"; "Whenever he heard her coming he would quickly put out his pipe".

2. Backshifted would. In a characteristic example of backshifted would, it corresponds to will in the complement of a past tense verb, e.g. "I will see them soon" vs. "I knew I would see them soon".

3. Modal remoteness. Huddleston distinguishes two subcases of modal remoteness that are relevant to would: remote conditionals and tentativeness.

Let's look at the tentative use first, since it corresponds to the notion of "doubt" that comes so easily to some people's minds in connection with would. This is a class of cases where "a rather vague element of tentativeness, diffidence, extra politeness, or the like" comes into play. Huddleston's examples include

He'd be about sixty.
Would you tell them we're here?
… and, I would suggest, it's too expensive anyway.
They would appear to have gone without us.

His comment on the last example is:

appear is a lexical modal of medium strength, qualifying my commitment to the truth of the modalised proposition: compare unmodalised They have gone without us. Would then adds further modal qualification, so that [the expression would appear] provides a double hedge against being wrong. […] The same effect is found in would seem, I would think, etc.

As Huddleston indicates, "tentative would" is a useful tool, and several particular versions of it are common. Perhaps these are the cases that lead some people to conclude that would is all about doubt. ("Tentativeness" seems to describe these examples better than "doubt" does, but the difference is a small one.)

However, as we'll see below, other uses of would, for instance those in what CGEL calls "remote conditionals", are much more common than "tentative would" is. At least, they're many times more frequent in all of the sorts of texts that I've looked at; and the divergence is especially great in the case of examples involving the fragment "That would be …".

In order to analyze these other uses of would, we need to go over CGEL's distinction between "remote" and "open" conditionals, discussed in section 14.2 of Chapter 8 ("The clause: adjuncts"). The basic idea:

The remote construction differs from the open in that it entertains the condition as being satisfied in a world which is potentially different from the actual world.

In other words, these examples deal with the evaluation of a hypothetical situation.  CGEL's terminology is better than the traditional "counterfactual conditional", because the world of a remote conditional is only potentially different from the actual world. And in fact, I think that the examples given in CGEL's treatment emphasis the unreal character of the remote-conditional world a bit too strongly:

If he tells her she will be furious. If he told her she would be furious.
If you are under 18 you need parental approval. If you were under 18 you would need parental approval.
If he bought it at that price, he got a bargain. If he had bought it at that price, he would have got a bargain.

In the conditional schema [If P, Q] CGEL explains that "… the remote conditional, like the open, excludes the case where P is true and Q false", but notes that remote conditions "where the time of the protasis situation is present or past", remote conditionals "generally implicate that P is false, or at least likely to be. And by the only-if implicature Q will likewise be false, or probably false."

This is why such conditionals have traditionally been called "counter-factual". However, the explanation continues,

It is important to emphasise … that a present or past time remote conditional does not entail that P is false (in the actual world); this is why we spoke above of the condition being satisfied in a world which is potentially different from the actual world. In the first place, it may well be that I don't know whether P is true or false … Secondly, there is one use of the remote construction, not common but nevertheless clearly-established, where I know or am confident that P is in fact true:

If he had escaped by jumping out of the window he would have left footprints in the flower-bed beneath. And that is precisely what happened.

At this point, I need to draw attention to a very common kind of if P Q construction, left out of the CGEL discussion, where Q has would (or another preterite-form modal auxiliary) but there is no presumption at all that P is false — and the if-clause typically has a plain present verb form. Some examples from the web:

If you want a domestic SUV, then the best choice would be the Ford Escape Hybrid.
If you want functional air filters, the best choice would have to be K&N Air Filter.
…if you really want to use WMA, the best choice would be 2-pass encoding.
If you really need a fighter bay, the best choice by far would be a manta bay.

Such sentences discuss a hypothetical situation, created by laying out a disjunction that divides up the space of circumstances affecting answer: If X, then A would be best; if Y, then B would be best; etc.

Very similar examples exist with present-tense forms of to be:

… if you want real wood, the best choice is engineered flooring.
If you want to stay in colonial Quito, the best choice is the Hotel Real Audiencia.
If you want HD programming, Dish Network is the best choice.

In these examples, the choice between is and would be doesn't seem to make a great deal of difference. Perhaps the preterite emphasizes the hypothetical nature of the situation under discussion — but with respect to perceived speaker (un)certainty, perhaps the preterite makes the speaker seem more rather than less certain, by emphasizing that the logic of the hypothetical situation is being carefully evaluated.

In any case, this same sort of situationally-partitioned assertion can be expressed without any explicit if-clause. This might involve replacing the if-clause with a prepositional phrase or some other way to reference the hypothetical state of affairs under consideration, as in these web examples:

In that case, the best choice by far would be the Ruger.
For VGA devices (640×480 screen) the best format would be XVID.
My recommendation in DLP would be the Mitsubishi HC1500.
In Standard Yiddish the word would be grubn.

And in other cases, the hypothetical situation — counter-factual or not — is set up in a question, and then assumed in the answer:

Q: I want to fly helios in the military what are the requirments and what is the best way to get there?
A: [the] best way definitely would be to go warrant officer in the army.

Q: What exercise is best for muscle 1 type fibers?
A: … the best way to train this type of muscle fiber would be with low intensity endurance type exercises

Q: What shampoo is best for oily hair?
A: The best one I've found so far would be Tea Tree from Paul Mitchel… Another way for your hair not to be so oily would be to not use conditioner …

Q: What would earth be like without a moon?
A: The most obvious difference would be that there would be no tides at shorelines.

And in other cases, the question is only implicit:

"The best solution would be for us all to become vegetarians".

So suggested the head of the UN climate agency, Yvo de Boer, who is attending UN-led climate talks in Germany this week. He was responding to criticism that measures to tackle climate change are partly to blame for the rise in food and energy costs.

De Boer's statement answers an implicit question, perhaps something like "What should we do to deal with the hypothetical (and also actual) situation that some allegedly green energy measures, like ethanol from corn, raise food prices?"

Or in President Obama's 3/24/2009 news conference:

Now, have — am I completely satisfied with all the work that needs to be done on deficits? No. That’s why I convened a fiscal responsibility summit, started in this room, to start looking at entitlements and to start looking at the big drivers of costs over the long term. Not all of those are reflected in our budget, partly because the savings we anticipate would be coming in years outside of the 10-year budget cycle that we’re talking about.

All of these uses of would seem to deal with hypothetical situations — or sometimes with actual situations that are being treated as if they were hypothetical, often as part of a process of matching answers to alternative questions, or solutions to alternative problems ("For someone who wants X, Y would be the best choice").. Thus it seems plausible to count them all as instances of "modal remoteness", even if they don't count literally as instances of either Huddleston's "remote conditional" or "tentativeness" categories. In the survey below, I'll call these "implicit remote conditionals", though it might be preferable instead to include remote conditionals in a broader "irrealis" category of reference to hypothetical situations.

In order to get a crude sense of the relative frequency of these different uses of would in recent American English, I used a random sample of 50 of the 891,081 instances of this word in the 385M words of the COCA corpus, respresenting various sorts of American English texts dating from 1990-2008. The results:

CGEL Category Description Count
(of 50)
1 (past time would) volition 1 2%
propensity 6 12%
2 (backshifted would) future will in past time 18 36%
3 (modal remoteness) tentativeness 1 2%
explicit remote conditional 6 12%
implicit remote conditional 18 36%

This sample had no examples of what I've called the TWBM construction.

Examples of backshifted would from my sample of 50 instances:

The government hoped at the time that this unusual step would both calm markets and lead to a recovery.
Mrs. Doncaster said Dad would take them in, but he wasn't back yet.
Air resistance was yanking her down, and in moments she would be tumbled to pieces on the ground.

Examples of explicit remote conditionals:

They're the ones who aided and abetted his bookishness. He would've been a normal kid if not for them.
And the second one would be, if you don't, would this enable science, maybe, not to use testing on animals anymore …
In fact, we would almost certainly be healthier if we did imitate Hollywood …

Examples of implicit remote conditionals:

Most families would be disappointed to have a gay son.
People who would otherwise think torture is wrong justify it because the world is evil.
Such a transition would mark the creation of social boundaries in addition to the symbolic boundaries marked by literacy and the pursuit of connections to the state.

From my point of view, the take-away message here is that cases of "modal remoteness" make up about half the total, and that discussions of hypothetical situations without any explicit if-clause — what I've called "implicit remote conditionals" — are about 3/4 of those.

Spot checks of other sources show the use of would similarly dominated by hypotheticals, just as Rodney Huddleston noted in the quote at the start of this post. Thus according to the transcript of President Obama's 3/24/2009 news conference, would was used 19 times, and 13 of the them were "implicit remote conditionals" like these:

And so this provision would affect about 1 percent of the American people. They would still get deductions. It’s just that they wouldn’t be able to write off 39 percent. In that sense, what it would do is it would equalize.

Among the 13, I haven't counted this one, which I counted as an instance of the TWBM construction:

Q: Isn’t that kind of debt exactly what you were talking about when you said passing on our problems to the next generation?
First of all, I suspect that some of those Republican critics have a short memory, because as I recall, I’m inheriting a $1.3 trillion deficit, annual deficit, from them. That would be point number one. [emphasis added]

But perhaps this should have been the 14th "implicit remote conditional": "If I were responding to those Republican critics, that would be point number one".

This suggests one plausible source for the TWBM construction, the one that  I had in mind when I suggested that it's used to present an "evaluative reaction to a hypothetical situation". At least, I think that this is half of the story. More on this later.


  1. Mr Punch said,

    April 3, 2009 @ 9:21 am

    Exactly. "That would be me" ordinarily means something like, "I am the person referred to, but I do not necessarily accept the truth of the statement."

  2. Amy Stoller said,

    April 3, 2009 @ 9:47 am

    "If he had bought it at that price, he would have got a bargain."

    Or, "… would have gotten a bargain," in American English.

    "Gotten" is so engrained in most Americans that it takes a lot of work before they stop "correcting" their characters' dialogue when "would have got" and so forth is used in English plays.

  3. acilius said,

    April 3, 2009 @ 10:32 am

    Thanks for these posts. I've long wondered about "That would be me."

    I have a vague idea about the construction that I could develop only if I knew a lot more about speech act theory than I do. In hopes that someone might be reading who has such knowledge and would be interested in helping me flesh it out, here's my vague notion.

    A modal like "would" calls attention to the act of speaking. Perhaps it calls attention to the tentativeness of the act, perhaps to some other circumstance of it. So, one who hears the exchange ("I'm looking for Liberman." "That's me.") might think, "So that's Liberman." One who hears ("I'm looking for Liberman." "That would be me.") might think, "So he's saying that he's Liberman."

    If "I'm looking for Liberman" is spoken with menace, there could be quite a difference between these responses. The bare statement "That's me" doesn't acknowledge anything special about the act of identifying oneself as Liberman. Perhaps it might be spoken by an interlocutor who missed the note of menace, or who is trying to ignore it. "That would be me" does- you're here to confront Liberman, eh? I'll not hide from you- do your worst! Perhaps also the reason "That would be me" sometimes raises a smile in a non-hostile context is just this, that it declares a willingness to identify oneself. If no one has anything against Liberman, it might be funny to make a point of your courageous willingness to own up to being Liberman.

  4. the Fish said,

    April 3, 2009 @ 10:44 am

    None of these situations seem to account for the usage in your original statement, "That would be me."

    The would in this statement seems to be for emphasis, rather than tentativeness. Not only, "It is me," but, "…as opposed to all other people who it could be." There also seems to be a sense of the fitness or inevitability of the answer, with a slight implication that it should have been obvious.

  5. Stephen Jones said,

    April 3, 2009 @ 10:45 am

    If he had escaped by jumping out of the window he would have left footprints in the flower-bed beneath. And that is precisely what happened

    But in this case the protasis is open.

    Where the past tense/second (unmarked) form is used to describe distance in time, then we have an open protasis. Take these two examples.
    If he leaves now, he'll be here by eight.
    If he left first thing this morning, he'll be here by eight

    In both cases we're describing a real possibility.

    In the following case we don't know if we are describing a real possibility or not.
    If he'd left at six o'clock he'd be here by eight.

    Where however the past/unmarked/second/distant form is describing something that is distant in reality, then we can't use 'will' in the apodosis, but must use a distant modal.
    *If I left first thing this morning, I'll be here by eight.

    And then there is the interesting case where 'would' is not implying distance but habit.
    If he walked to work, he would always wave to the children on the other side of the road going to school.

  6. marie-lucie said,

    April 3, 2009 @ 11:36 am

    I think that "habit" is more to the point than "propensity" for use of would as in SJ's last example. "Propensity" to me suggests a value judgment that the act after would is something to be avoided, but in most cases that use of would simply implies a typical, habitual activity of the subject with no value judgment attached. As with any habit, there might be the occasional break in the occurrence of the activity, but it considered as generally true.

  7. Tristan said,

    April 3, 2009 @ 12:55 pm

    It seems like TWBM most closely resembles the Q&A-style implicit remote conditional.

    A: I'm looking for Mark. [Where's Mark?]
    B: That would be me. / *I would be that. / (?) I would be him

  8. Sky Onosson said,

    April 3, 2009 @ 12:56 pm

    I think there is a way to bridge the gap between use of would to indicate tentativeness, and its use to imply certainty. Both can be summed up as "drawing a conclusion". I am not sure if this can be related to all senses of would (nor am I suggesting that there should be on catch-all meaning), but I think this does cover quite a few cases: backshifting, tentativeness, remote/counterfactual (implicit and explicit). In fact, the explicit remote conditionals described in the post lay bare the conclusion-drawing function of would, I would think! the Fish's point about TWBM implying inevitability and obviousness also fits quite well with this, too.

  9. Sky Onosson said,

    April 3, 2009 @ 12:57 pm

    edit to me previous post: "on catch-all meaning" should be "one catch-all meaning"

  10. Sky Onosson said,

    April 3, 2009 @ 12:58 pm

    *argh* edit: me=my

    (distant grumbling about LL not having comment-editing…)

  11. Bob Zurncle said,

    April 3, 2009 @ 1:19 pm

    "distant grumbling" . . . I like that. It picks up some of the feeling of "a distant rumbling", and crosses it with a sort of dimunition of the act of grumbling: I distance myself from the idea of grumbling as I don't take said grumbling all that seriously. And still, it adds the sense of grumbling with a dark cloud over one's head, as with the distant sound of thunder. Neat.

  12. comwave said,

    April 3, 2009 @ 1:23 pm

    I think "the speaker's level of information on the questioner's information" can be a guideline in the distinction and an answer to the question of why TWBM construction.

    "That is me."
    When the speaker believes h/she has FULL information on the amunt of information the questioner has regarding the question.

    "That would be me."
    When the speaker believes he/she has INCOMPLETE information on the amount of information the questioner has regarding the question.

    There are many levels of incompleteness depedning on situations. I think efforts to group those levels into a few categories in the study of modals have been made with no consideration of information level. (This judgment of mine should be retreated had there been one.)

  13. marie-lucie said,

    April 3, 2009 @ 1:40 pm

    SO: (distant grumbling about LL not having comment-editing…)

    What do you mean? the gray space under "submit comment", activated when you start to write, shows you the exact format in which the comment will appear, more legibly than in the writing window. You can edit in the writing window and see the edited result in the gray window. What else do you need?

  14. bianca steele said,

    April 3, 2009 @ 1:54 pm

    The post above suggests that "would be" constructions are perceived as expressing personal uncertainty because they can be hypothetical ("if she is under 17, she would not be able to buy a ticket to this movie"). This could be tested by comparing with reactions to a complete syllogism ("if she is under 17, she would not be able to buy a ticket to this movie, and she is 14, so she will not be able to buy a ticket"). The latter is explicit about the speaker's having a complete familiarity with the situation, both in theory and in practice, and leaves no loopholes for a counterargument, and therefore is more certain.

  15. dr pepper said,

    April 3, 2009 @ 2:00 pm

    I just thought of another construction:

    "Where's Liberman?" "How would i know?"


    "So, i hear Liberman is planning to write a novel entirely in snowclones." "I wouldn't know."

  16. Clarissa said,

    April 3, 2009 @ 2:56 pm

    This usage (distancing/politeness/tentativeness) is extraordinarily difficult to teach to English learners, partly because I don't fully understand when and why I use it myself. More discussion of it would (…) be really welcome!

  17. D.O. said,

    April 3, 2009 @ 4:14 pm

    I might be really dense, but it seems to me that 2 of your 3 examples from COCA for backshifted might actually be from the "remote" class.
    "The government hoped at the time that this unusual step would both calm markets and lead to a recovery." almost cries for "and they were wrong" or at least "and they were right". Of course, it might be a usual future-in-the-past, but "hoped at the time" strongly suggests otherwise (do I have my F already?).
    "Mrs. Doncaster said Dad would take them in, but he wasn't back yet." This is also not very clear. What did Mrs. Doncaster really said "Dad will take you in", but then it turned out that dad wasn't back yet or she herself said "Dad would take you in, but he is not back yet" ? Is the last option overruled by "would have taken them" if this is really what she meant? Tell me I am a really ignorant ESL guy.

  18. Sky Onosson said,

    April 3, 2009 @ 5:57 pm

    ML: You can edit in the writing window and see the edited result in the gray window. What else do you need?

    Well, where do I begin?

    In all seriousness, though, BZ was correct in that I was trying to distance myself from complaining too loudly. While those of us (myself included) who are not always so swift might find it useful to be able to retroactively edit our comments, I also realize that there are good reasons to not allow this.

  19. marie-lucie said,

    April 3, 2009 @ 8:58 pm

    SO: Well, where do I begin?

    Obviously you know where to write a comment, but you probably haven't gone down to the bottom of the page. Bring the cursor (right hand side of the screen) down to the very bottom of the slide, so you can see "submit comment". As soon as you start writing, the gray window will appear below "submit comment", and the letters you type will show up at the same time in the writing window and in the gray window. You can then see what have written more clearly, as it would appear after submitting your comment, and you can correct or edit it in the writing window as with any word processor. (Don't try to write in the gray window).

  20. Sky Onosson said,

    April 3, 2009 @ 10:02 pm

    I think you misunderstand, ML, that comment was meant to be a bit tongue-in-cheek (i.e. the list of things I need is immense, due to my many deficiencies), but I should really know better than to attempt that on the internet…

    I definitely know where the preview text is below the comments – but sometimes, even the best of us miss something when editing, and a stray "e" or a "me" instead of "my" can often slip by unnoticed… until the comment goes live and it sticks out like a sore thumb.

    Anyways, I will try to be more careful in the future, both in my previewing and my attempts at humour!

  21. marie-lucie said,

    April 3, 2009 @ 10:40 pm

    SO, OK, it would not be the first time I was told I had missed the humour in someone's comments. But I had not noticed the second window myself until some time after I started posting comments. It is true that some typos still slip by unnoticed until finally posted, but seeing the second window does reduce the number of such errors.

  22. boynamedsue said,

    April 4, 2009 @ 7:01 am

    The ((if subj + present simple), (subj + would + be)) constructions that you show all involve situations of recommendation, and most contain the word "best".

    e.g.> "Q: I want to fly helios in the military what are the requirments and what is the best way to get there?
    A: [the] best way definitely would be to go warrant officer in the army."

    These are a set construction "The best x (to y), would be (to Z)". In EFL these constructions are taught separately to other conditionals (usually at Advanced C1 level).

    The examples that don't use best are buried present hypothetical conditionals.

    >"In Standard Yiddish the word would be grubn" = "If this word was standard Yiddish (which it isn't), it would be Grubn

    >"Q: What would earth be like without a moon?" =
    If the earth didn't have a moon what would it be like?

  23. Shaunta Borquez said,

    March 27, 2014 @ 12:29 pm

    Hi, all the time i used to check web site posts here early in the break of day, because i enjoy to gain knowledge of more and more.

RSS feed for comments on this post