Monday, August 11, 2014

Nonpropositional views of conditionals, and lying

  1. Some conditional claims are lies. ("If you show this car to any mechanic in town, he'll tell you it's a great deal.")
  2. Conditional claims do not express propositions (but, say, conditional probabilities).
  3. Assertions express propositions.
  4. So, not all lies are assertions.

Of course, (2) is a quite controversial theory of conditionals. And one can turn the argument around: All lies are assertions, so conditional claims express propositions (at least in those cases; but one can generalize from them). But if one thinks that the argument should be turned around in this way, then one must make the same move for every non-cognitivist theory, since it takes only one non-cognitivist theory in whose domain lies can be made to yield conclusion (4). For instance, one must reject non-cognitivism in metaethics and aesthetics. So far so good. One probably doesn't need to reject non-cognitivism about requests, on the other hand, since one doesn't lie with requests: "I'll have fries with that" isn't a lie when you don't want fries.

Are there domains where one can lie but where non-cognitivism is clearly right? I am not sure. Maybe something like talk of the spooky? One certainly can lie in something is spooky (e.g., to scare off a competing house buyer). But even there I am not convinced that the right conclusion is that not all lies are assertions. The right conclusion there may still be that we do in fact make assertions when we attribute spookiness.

Vagueness cases are another case to think about. I think propositions are always sharp, but we lie with claims that clearly do not express sharp propositions ("He was bald two days ago, but washed his hair with this shampoo, and now he's not"). But I don't think vagueness cases give one reason to accept (4). Rather, they lead to a refinement of (3): assertions don't express individual propositions, but something like a vague assignment of weights to propositions.

So overall, I don't know what to make of the argument.


Mike Almeida said...

Is the initial conditional a lie? It says that you won't show this to any mechanic in town or he will tell you its a great deal. First, it might just be a prediction. I don't think predictions can be lies. It might be a promise, as in "I promise you that he will say X". That's not a lie either, even if you know he won't say X. The conditional is a lie, I suppose, only if you know that I will show this car to some mechanic and that he will not tell me it's a great deal. But even that is not obviously a lie. If I show the car to some mechanic and the mechanic is overwhelmed with the car's wonderful features and instantly dies, I don't think you lied to me. That is, despite the fact that he did not tell me the car was a great deal.

Alexander R Pruss said...

Surely it's not a material conditional.

Mike Almeida said...

If you say so, then it's so. It's your conditional. But the main objections do not depend on it being essentially a disjunction. It might still be a prediction or a promise. And it is not obviously a lie even if the antecedent is true and the consequent false. So, it is sort of fascinating to wonder when a conditional is a lie.

Alexander R Pruss said...

I don't think "I promise that 2+2=4" is a promise. It's a way of definitively asserting that 2+2=4. Promises always concern (directly or indirectly) one's own future behavior. (If I promise that my child will do something, I am really promising to ensure they do it.)

So, I don't think this is a candidate for a promise.

I am intrigued that you don't think predictions are assertions. Hedged predictions may not be assertions. But predictions stated definitively seem to be just future-tensed assertions.

Mike Almeida said...

I think I said that I don't think a prediction can be a lie. If I predict that you'll get a raise this year and you don't, I didn't lie to you. On promising, it is certainly possible for me to promise that you'll receipt from my company, though I'm not the one proving the receipt. In this case, I promise concerning the behavior of others. I can make promises of all sorts, depending on what I know. I can promise you that the train will arrive on time, given my knowledge of the schedules and the condition of the train, etc. I don't see anything unusual in that.

Brett Lunn said...

Suppose God knows the future and can reveal to me so as to constitute knowledge given the truth of the claim that the stock market will rise 25% in the next week. Next, say I tell a stock broker i know (and apparently dislike) who knows that God revealed to me the future of the stock market that the stock market will drop 10% in the next week. This seems to be both a prediction and a lie, but let's look at some points of attack.

First, one could deny foreknowledge. I don't care to argue for that here.

Second, one could say this scenario utilizes middle knowledge since it doesn't fulfill Pruss' relevant similarity condition. However, I think we can tailor my rough example to fulfill the relevant similarity condition.

Third, one could say it isn't lying. However, it seems to be a counterexample to the most common definition (source: SEP). I also think it probably can be tailored to fit most, if not all, definitions.

Fourth, one could say it isn't a prediction, but I'm not sure how one would go about doing that? Must a prediction entail that it isn't knowledge or something? I'm not sure how to pursue this line.

Maybe I missed an objection, so I'm looking forward to your response.

Mike Almeida said...

That's not a prediction. I can't predict ~p when I know that p.

Brett Lunn said...

Why not? What exactly is a prediction, then?

Alexander R Pruss said...


This is a tricky case. I don't like causal loops, and here we have something like a causal loop, in that God's prediction is presumably caused (or quasi-caused) by future events, but then God's prediction causally affects the future events. I think prophecies that are predictions (rather than promises) may need to be causally isolated from what is predicted. Illustration: When Jesus predicts Peter will deny him, Peter forgets the prediction; the forgetting is the causal isolation.

Alexander R Pruss said...


"I think I said that I don't think a prediction can be a lie."

I am now inclined to agree with you, but for a very different reason. A prediction can't be a lie because a prediction can't be insincere. Just as knowledge is factive, prediction is sincere. Suppose that I am a crooked financial analyst and I want Apple stock, which I have a lot of, to rise. With no evidence, I tell big clients that Samsung will declare bankruptcy next month. It would be incorrect to say that I predicted that Samsung will declare bankruptcy, because it wasn't a prediction, but just a pretend-prediction.

Mike Almeida said...

In the case you describe, there are true, future contingent propositions as well as true present and past contingents. Suppose P is a present contingent truth, say, it is raining outside. I utter the sentence 'i predict it is not raining outside', knowing that it is raining. Did I lie to you? No. Suppose Q is a true future contingent, say, you will get a raise tomorrow. I utter the sentence 'I predict you will not get a raise tomorrow'. Did I lie to you? No. In your case, too, I do not lie to you. I utter 'I predict the market will decline'. I do not utter 'the market will decline', so I do not so much as say anything false, let alone lie.

Alexander R Pruss said...

Well, if you don't predict that p, then when you say "I predict that p", you do lie.

And of course in general the truth and falsity of what you assert has nothing to do with whether you lie when you make the assertion.

Brett Lunn said...

Dr. Pruss,
Well I think we can tailor the example so that the stock broker forgets, has his memory wiped, or dies instantly. This would prevent the causal loop.

Also, maybe we can find a case similar to your examples in sincerely asserting what one does not believe. Then make that case into a prediction and maybe we have a lie. However, that wouldn't exactly be easy.

Dr. Almeida,
Suppose a person knows it is raining outside and predicts that it isn't. Also, suppose that person knows that if he predicts something, then the person they are talking to will believe that thing is true (say, a parent talking to their child). Given that, I tend to think the person is lying. (Or maybe Dr. Pruss' point is more clear cut.) But I'm open to be persuaded.

Mike Almeida said...

Really interesting case, Alex. Suppose I utter 'I predict that p' and I don't thereby predict that p. I'm not sure how that might happen, but let's just say that it might happen. If so, I don't lie, since my utterance 'I predict that p' does not express the proposition :I am predicting that p: or :it is the case that I predict that p:. These propositions can be true or false. What I express by the utterance 'I predict that p' cannot be true or false, and so cannot be a lie.

Mike Almeida said...


If I utter 'I predict it is not raining' in the circumstances you describe, then I'd say that I'm misleading you or maybe even that I am manipulating you, but not that I am lying. I think to utter a lie I have to utter something that I believe is false, but my prediction is not the sort of thing that can be false (or true)

Heath White said...

I do not know why anyone would put a sincerity condition on predictions. Pundits make market and political predictions all the time (at least this is what journalists and media types call them—just google “market prediction”) and who knows whether they are sincere or are just filling column inches and airtime. But who cares? It’s an assertion about the future which does not depend on the choices of the assertor. That is what I would call a prediction.

If there can be knowledge of the future that does not depend on one’s choices, then one can lie about it. If I tell a two-year-old that the sun will not rise tomorrow, because God is very angry with his misbehavior, I am both predicting and lying. (And not necessarily lying about God being angry.) Also, I would be lying if I put this statement in conditional form: “And if you keep on misbehaving, the sun won’t rise the next day either.”

Brett Lunn said...

Dr. Almeida,
Supposing that I cannot (at least, usually) rationally predict both p and ~p, then combining my example with Dr. Pruss' point seems to entail that the person is lying, no? But I suppose you would reiterate your response to Dr. Pruss' point, so one would need to settle that first.

Anyway, the internet on my phone keeps crashing on this site for some reason, so thanks to both for the exchange.

Mike Almeida said...

Heath and Brett,

Just to avoid all of the ambiguity, let's make your prediction explicit. Suppose you utter this sentence:

P = 'I predict that the sun will not rise tomorrow because God does not want it to'

P is a prediction, and it does not have a truth value. It is not false, so you could not believe that it is false and utter it. So, it cannot be a lie.

I agree that there might well be some other moral fault in your uttering P. It is misleading, etc. But it just isn't a lie. A minimal condition on a lie is that it is something you believe is false.

Heath White said...

There is a literature on this with respect to assertion, as in what to say about the truth values of sentences like “I assert that I have never been in Wyoming.” I think, and I believe the philosophical consensus and the law back me up, that if you are asked under oath whether you have ever been in Wyoming and you reply “I assert that I have never been in Wyoming,” and you have been in Wyoming, you have perjured yourself, i.e. lied under oath.

There is, of course, another sense in which statements of the form “I assert that p” are self-verifying. After all, you ARE asserting p. What to say about this?

My preferred position on this is that truth values go with speech acts, and there are two in this case: one assertion of “I have never been to Wyoming” which is false, and one assertion of “I assert that I have never been to Wyoming” which is true. The perjury is due to the first act not the second.

I would say the same thing about predictions, which I think of as a species of assertion (i.e. about the future, not under one’s control). The sentence “I predict that p” asserts both that p, and that one predicts that p. (Mike’s no-truth-value stance is odd, since it implies that ‘I predict p’ in my mouth and ‘Heath predicts p’ in your mouth express different propositions.)

But even if you don’t like that view, it seems very clear that if the television interviewer asks, “Do you predict a market drop?” and I answer, “I predict the market will drop tomorrow,” I have answered his question, and therefore made an assertion.

Also, the speech of prediction (like that of assertion) does not need to be explicit. “The market will drop tomorrow” is a prediction (and an assertion) too. It has a truth value, expresses a proposition, and can be a lie.

Mike Almeida said...

But even if you don’t like that view, it seems very clear that if the television interviewer asks, “Do you predict a market drop?” and I answer, “I predict the market will drop tomorrow,” I have answered his question, and therefore made an assertion.

I agree that that is an assertion you are making. But that's only because your using a manner of speaking to assert this 'yes, I predict p'. I never denied that assertions of that sort have truth values. Nor did I ever deny that sentence of the sort 'I assert that p' express propositions that have truth values. I think I said nothing about such sentences.

What I did say is that utterances of sentences of the form 'I predict that p' do not express propositions that have a truth value any more than do sentences of the form 'I promise that p'. That is what explains how bizarre it is to ask the following questions:

1. I predict that it will rain

2. Is that true?

Is what true? I didn't say anything that might be true or false. You might be asking "is is true that you so predict", but I did not say in (1) that I so predict. I say that in (1').

1'. I predicted that it will rain or I often predict that it will rain or I am now predicting that it will rain, etc.

Predicting that p (as in (1)) is not saying that I predict p (as in (1')). Those are two different propositions, and I'm pretty sure they're being conflated here. Now to promising:

3. I promise to mow the lawn.

4. Is that true?

Is what true? That I promised? Yes, it is true that I promised, but I did not say that I promised anything. I do that in (3').

3'. I promised to mow that lawn or I often promise to mow the lawn or...

Making a promise that p (as in in (3)) is different from saying that I promised that p (as in (3')). These two propositions are being conflated.

Alexander R Pruss said...


I'm not sure about "I assert ..." It seems to me that "I assert that p" is nothing but an assertion that p. It's just an accident of our language that we don't routinely prefix assertions with the performative "I assert that" in the way that we prefix promises with "I promise that". There is no second meta-level assertion being made.

Likewise, when you say "I promise that p", you are making a single speech act, a promise that p, rather than two speech acts, a promise and an assertion about the promise. We don't, I think, want to multiply speech acts beyond necessity. In a context where "I shall stand by you through thick and thin" makes a promise, it is a speech act pretty much the same as "I promise to stand by you through thick and thin." Clearly, "I shall stand by you through thick and thin" does not contain any assertion about promises. Thus, "I promise to stand by..." also does not contain any assertion about promises.

If this is right, then "I predict that p" also is not an assertion about myself making a prediction. It is simply a prediction.

But a prediction is not, I think, a kind of assertion. It is an assertion-like act, but with weaker evidential norms than assertion proper (and with a restriction of the subject matter to something futurish): I think you get to say "I predict that p" sometimes when you don't have enough evidence to say "p".

By the way, in the parentheses I vaguely wrote "something futurish", because I just imagined a case like this. Your friend just overate. You say: "I predict you will have a stomach ache in an hour." But you also predict that in ten minutes your friend will step into a time machine set for the the distant. So you're predicting an event that according to our generally shared timeline, and according to your own timeline, is past. It's future, though, according to your friend's timeline, and I suppose that's what counts. But it's going to be hard to specify just which timeline the prediction needs to be a prediction with respect to.

Heath White said...

I agree that one ought not multiply speech acts beyond necessity. But I think it’s necessary!

Two arguments:

1. If I say, “I assert that my coffee is cold” and you overhear this and tell someone else, “Heath asserted that his coffee is cold,” it sure seems like your statement and my statement have exactly the same truth conditions. The only thing that changed is the pronouns. But if that is right, then I have made a statement about my assertion and not merely about my coffee.

Similar remarks go for “I warn you not to park there” vs. “Heath warned me not to park there” and “I order you to stand still” vs. “Heath ordered me to stand still.” In each case, the first-person statements are not mere warnings and orders; they also say that one is warning and ordering.

2. Suppose I am captured by the enemy and they allow me to send a message home: “I assert freely and under no compulsion that I have not been tortured.” This means more than “I have not been tortured” since it includes the content that I am asserting this freely and without compulsion. (Of course, that is the very thing that people might not believe.) So this is a statement about my act of assertion. And similarly for the simpler, “I assert that…”.

Alexander R Pruss said...


Do you want to say the same thing about "I promise that..."? For the third-party report argument can be equally made about promises.

I agree that when you add adverbs, you get an assertion. Likewise for promises: "I freely promise..." contains the assertion that I promise as well as the promise. But I don't think "I promise..." contains an assertion.

Still, you might be right. Here's a test case. Sally plans to torture Sam. However, she also believes it is metaphysically impossible to promise something wrong (a promise to do wrong is null and void, and her theory of null and void promises is that they aren't promises, just as fake money isn't money) and she believes that torturing Sam would be wrong. She says to Sam: "I hereby promise to torture you." She has every intention of torturing Sam. But she thinks she isn't making a promise.

Is Sally lying?

I don't know.

Heath White said...

I do want to say the same thing about ‘I promise.’ It seems a little odd to draw attention to the assertion that one is promising, since it is rather difficult for such speech acts to be infelicitous. (Same for prediction, assertion, warning, etc.) So they are self-verifying under normal circumstances. But the assertion is there, I want to say.

Your position on the adverb-enhanced prefix is odd, in that you want to say
(1) I freely promise that p
Entails … what? The obvious thing to say is
(2) I promise that p
But you also want to say that assertions of (2) do not in fact assert that I promise anything. So how do you frame the proposition which is, so to speak, (1) minus ‘freely’?

Your Sally case is tough, because (i) she is making a promise, which she plans to fulfill, but (ii) she thinks she is not making a promise. So if it is a lie, it is because it is an assertion with intent to deceive but not an actual falsehood. And the thing she would be deceiving him about is about a particular assertion that she is making (to be distinguished from a particular promise she is making), which is hardly the most salient issue between them. But it may not even be an attempt to deceive, if Sam’s beliefs are odd enough and she knows this.

So my thought is that yes it probably is a lie, but I don’t know how to make that clear.

Alexander R Pruss said...

I say that (1) entails the proposition that (2) would express in the odd contexts in which it was also an assertion. I just think in ordinary contexts, no assertion is made by utterances of (2).

There are presumably other sentences that in ordinary contexts are not assertions, but that could be used to make assertions. Certain exclamations, for instance.