1. “It’s 3 o’clock” in linguistics and in philosophy; 2. A “truism”?; 3. An Argument for “hard-ass” semantics

1. I remember discussing the opening of Peter Lasersohn’s “Pragmatic Halos” (Language 75 [1999]: 522-551) with a linguist friend of mine a few years back (email records show this was in March of ’05).  Lasersohn gives this example:

Suppose, for example, I tell John that Mary arrived at three o’clock.  In certain relatively unusual circumstances, the exact second of her arrival might be important, but most of the time this level of precision is not required.  So if John finds out later that Mary didn’t arrive at three but at fifteen seconds after three, it would be unreasonable of him to complain ‘You said she came at three!’

But whether or not John is acting unreasonably in this situation, I think we have to concede that he is, strictly speaking RIGHT: when I told him that Mary arrived, I said something that was literally false, not true.

I have some complaints about how Lasersohn presents this, but will let them pass for now, because I’d like to begin by just getting a handle (hopefully, with your help) on what folks think about this kind of case.

My linguist friend, as I recall, was on-board with Lasersohn here.  Me, not so much.  And as luck would have it, I happened to have just been reading a section of John Hawthorne’s (then, pretty recent) Knowledge and Lotteries, where John, as I am inclined to, issued the opposite judgment about the truth-value of such temporal claims, made in relaxed contexts:

Consider by analogy

(3) It is three o’clock.

When said at 3:01 (in the context of a rocket launch, or to someone running to reach an airplane whose doors close at precisely 3:00), (3) clearly expresses a falsehood. But there are plenty of ordinary, less demanding contexts, where intuitively (3) expresses a truth.  (p. 99)

It’s interesting that John uses the same case, with the same time, that Lasersohn had used a  few years earlier–which is unlikely to be a coincidence–while issuing the opposite verdict from Lasersohn about the truth-value of this kind of claims made in more relaxed contexts. What’s really interesting is not just that these two disagree about this matter, but that both take their verdict about this matter to be a starting point of their arguments, not anything controversial enough to need to be argued for before it can be used.  (John is not just reporting what he finds his intuition to be about this case only to question whether it is right: What happens next is that he uses his verdict that such claims are true in the argument he is giving.)

Since Larry the linguist was agreeing with the linguist, and Keith the philosopher was agreeing with the philosopher, a natural conjecture was that this was something that linguists and philosophers just tended to disagree about.  But while there may be something to this conjecture, I don’t think it really gets things right.  It seems to me that there are pockets of linguists/philosophers who think one way, and pockets that think the other (and philosophers may be more inclined than linguists to think the relevant claims are true).  And what’s curious is that at least some in each camp apparently think their verdict is uncontroversial enough to be used as the starting point in an argument, without first defending it.

So, what I’m wondering is: How popular are these two different takes on the truth-values of the relevant claims, in philosophy and in linguistics?  Is either answer popular enough (in one or the other discipline, or in general) to be at least relatively uncontroversial?

2. Lasersohn seems to think the position he takes on the truth-value of the relevant claims (those made in relaxed contexts) is a “truism.”  The very fist sentence of his paper’s abstract is:

It is a truism that people speak “loosely”–that is, that they often say things that we can recognize not be true, but which come close enough to the truth for practical purposes.

And I think the opening of the body of  the paper is written in such a way as to at least hint that Lasersohn is thinking that his 3 o’clock example is just a particular truistic (if I may) example of this.  (Though I imagine it is possible to read him as only saying it’s a truism that we often speak loosely in the sense he specifies, without  it being a truism that this particular example is an example of such loose speech.  Lasersohn’s paper is on JSTOR here, for those who’d like a look, and even those who don’t have JSTOR subscriber access can get the first page, which is what I’m discussing here.)

While this “truism” strikes me as pretty clearly false, and so not really a truism at all, I do think there’s a truism in the vicinity.  Indeed, even I, who think the relevant claims are true, not only think that we do in general engage in a lot of loose talk, but that the claims in question (claims like “It’s 3 o’clock,” made in relaxed contexts) are examples of loose speech.  I just don’t think we have to understand “loose” speaking according to analysis Lasersohn gives.  Supposing the kind of contextualist account of “It’s 3 o’clock” and the like that I’m inclined toward is correct, the claims in question are governed by relatively lax truth-conditions, and when a claim uses such a phrase (one that could instead be used in accordance with much higher standards of precision) in such a way, that’s speaking (relatively) loosely.

What is a truism about the claims in question is that their “acceptability conditions” are much more lax than are the acceptability conditions of claims made using the same words in other, more demanding contexts (and so they are in that sense examples of “loose speaking”).  About that Lasersohn and Hawthorne — and I — agree.  And it really does seem to deserve to be called a “truism,” I think (at least given a suitable explanation of what we mean by “acceptability conditions,” or the like).  But that such claims (claims like that “It’s three o’clock” made in relaxed context) are false?  That is very far indeed from being a “truism.”  Or so I think & hope!


3. Finally, there is an argument that I’ve quite frequently encountered (mostly in conversation), in favor of the “hard-ass” position on our question (that the relevant claims are all or almost all false, since they are invariably governed by very demanding truth-conditions, even if they somehow have quite lax acceptability-conditions).  I don’t think this argument is very strong at all, but what I’m wondering here is if anyone out there has a good sense for how influential it is among those who take the hard-ass position.  The basic idea of this argument is that, in just about any context, claims such as

a. It is three o’clock, but it’s actually slightly later than 3 o’clock

are pretty clearly wrong, as are claims such as

b. It is exactly three o’clock, but it’s actually slightly later than 3 o’clock,

while claims like

c. It is approximately three o’clock, but it’s actually slightly later than three o’clock


d. It’s three-ish, but it’s actually slightly later than three o’clock

sound perfectly fine.  This is often (with lightning speed, in my experience!) taken to “show” (the word I find is often used here) that, as far as truth-conditions about sentences like “It’s three o’clock” go, the hard-ass position has things right.

So, to what extent is this argument really driving the hard-ass position?  Are there other arguments that are influential here?  Or is it largely just driven by an intuition that such claims are false?  Any help would be appreciated.


1. “It’s 3 o’clock” in linguistics and in philosophy; 2. A “truism”?; 3. An Argument for “hard-ass” semantics — 3 Comments

  1. If we think about the utterance meaning/sentence meaning distinction, is there maybe not even a tension between Hawthorne and Lasersohn here? Lasersohn is addressing the meaning of the sentence, but when Hawthorne writes that “there are plenty of ordinary, less demanding contexts, where intuitively (3) expresses a truth”, we should take it that it’s the _utterance_ of (3) that expresses the relevant truth? (I haven’t gone back to look at John’s text, so I don’t know whether that hypothesis about what’s going on there is consistent with the rest of what is said in that passage.)

  2. I’m no expert on pragmatics or the like, but here’s an intuitively correct (i.e. I thnk it’s right) take on ordinary claims about the time: they are precise to the minute given. So “it’s 3 o’clock” means that it is after 2:59:59, but before 3:01. I think that with this account in hand, we can handle all sorts of cases. In ordinary cases, if you say to John that Mary arrived at 3 o’clock, and it was 3:00:15, your claim is literally true. But if instead she arrived at 3:01, your claim would be literally false. In cases that require greater precision like the space launch case, seconds count, so it’s not an ordinary case. In that case, “it’s 3 o’clock” would be literally false, if it was 3:00:15, because in that case “it’s 3 o’clock” is elliptical for “it’s 3:00:00.”

    (This account doesn’t appeal to the speaker/sentence meaning distinction, since it says that “it’s 3 o’clock” can be literally true, even if it’s 3:00:15. The account appeals, rather, to the view that a given context can allow a sentence to express a proposition that is more or less precise than the sentence appears on its face.)

    That doesn’t help with your questions, but perhaps it can offer a mediating viewpoint.

  3. Pat: I think something like the reading you explain is right for a lot of ordinary uses. (Though I suspect ordinary uses typically allow for rounding up, so that “It’s 3 o’clock” comes out true at 2:59:59, as well as at 3:00:08.)

    I would think it implausible, though, to hold that those are the truth-conditions for all standard uses of the likes of “it’s 3 o’clock.” And what you say about the cases requiring greater precision indicates that we agree on that.

    In the post, I mention, but don’t explain, “the kind of contextualist account of ‘It’s 3 o’clock’ and the like that I’m inclined toward.” I suppose that that’s the kind of general approach you’re inclined toward: One on which the truth-conditions vary among contexts (with some contexts demanding more precision than others), with something like the account you explain (though perhaps modified to allow some rounding up) giving the truth-conditions for lots of ordinary uses.

Leave a Reply

Your email address will not be published. Required fields are marked *