Lying is an important social and moral category. We react negatively to liars and their lies. But what is it to lie? The standard view in philosophy and social science is that a lie is a dishonest assertion. This view goes all the way back to at least the 4th century, when Augustine wrote, “He may say a true thing and yet lie, if he thinks it to be false and utters it for true.” On this view, lying is a purely psychological act: it does not require your assertion to be objectively false, only that you believe it is false.
About two years ago, my son Angelo came across an expression of the standard view of lying. He wondered whether it fit the ordinary concept of lying. (You might be able to imagine the sort of dinnertime conversations that could lead a twelve-year-old to become curious on this point.) In particular, Angelo was interested in whether, on the ordinary view, lying was a purely psychological act. So we conducted some behavioral experiments to find out.
Here is a pair of examples we used to test people’s judgments. In each version of the story, a man makes an assertion with the intention of deceiving his audience. The two stories differed in one critical respect: the man says something false in the False version, whereas he ends up saying something true in the True version.
[False/True] Jacob’s friend Mary recently posted information on the internet that will alert the public to serious government corruption. Soon some federal agents visit Jacob and ask where Mary is, in order to detain her. Jacob thinks that Mary is at [the grocery store/her brother’s house], so he tells the agents, “She is at [her brother’s house/the grocery store].” In fact, Mary is at the grocery store.
Interestingly, we found evidence that the response options offered to people had a dramatic effect on how they described the case. For example, if you force people to choose between saying that the man “lied” or “didn’t lie,” then they will say the man lied, even when he says something true. However, Angelo and I wondered whether this was an artifact of the mode of questioning and, thus, should not be taken at face value, in either of two ways. On the one hand, we wondered whether it was because people simply answered in accordance with how things seem to the speaker himself (perspective-taking). On the other hand, we wondered whether people thought that answering “didn’t lie” could misleadingly suggest that the man was not trying to lie, or that he did not deserve blame for his dishonest intentions (pragmatic implication).
Accordingly, we conducted two follow-up experiments that gave people more flexible response options. These options disentangled the agent’s perspective and intentions, on the one hand, and the attribution of lying, on the other. If people continued to classify the man as a liar in the True condition, then it would show that the standard philosophical view nicely captures the ordinary concept of lying on this point. By contrast, if people no longer classified the man as a liar in the True condition, then it would show that the standard philosophical view and the ordinary concept differ fundamentally on this point.
As it turns out, when given more flexible options, people no longer classified the man as a liar in the True condition. Instead, they judged that he tried to lie but failed to do so. Alternatively, they judged that he only thinks he lied. Here is a particularly striking figure from our final study, in which people selected between “He tried to lie and actually did lie” and “He tried to lie but only thinks he lied.” The bars represent the percent of people in the two conditions who said that the man did lie.
Overall, we concluded that, on the ordinary concept, lying is not a purely psychological act. As an added bonus, in the process, Angelo might have become the youngest first-author ever on a paper published in Cognition!
[Cross posted at Experimental Philosophy]
Interesting.
Have you read Roy Sorensen’s work on lying (and affiliated notions)? He didn’t do any fieldwork, but you might be interested in it anyway. Or Angelo might be.
“Lying with Conditionals”, upshot: you can lie with a false conventional implicature, but not with a false conversational implicature. “Bald-face lies!”, why are lies that do not and could not deceive anyone wrong? “Knowledge lies”, in which the deception attempts to undermine the audience’s knowledge but not by implanting a false belief. And I think there are one or two others.
Thanks, Jamie! That work definitely sounds interesting and, because it’s Sorensen, probably really insightful. The point about lying with conditionals sounds like it could be another example of lying with true statement (true conditional asserted + false conventional implicature = lie).
Good for Angelo, what a young achiever! This is all interesting, but I’m not yet convinced the results tell us much either about what lying is or even about what “the ordinary concept” of lying is. The phrase “He tried to lie and actually did lie” is a mighty unusual one, and this and the other choice of response offered might be taken by respondents to have their own pragmatic implications.
I’d be curious about an analogous experiment that might bring this out. The ordinary meaning of “shot at” is something like “shot in the approximate direction of”. Does the ordinary concept of “shot at” require the shooter to have actually hit the target? Obviously not; it is possible to be shot at without being hit. Here’s the case, then:
Jacob looked through the sight, aimed the rifle at Mary and pulled the trigger. Jacob thought the bullet hit Mary. In fact, the bullet [struck Mary’s arm / embedded itself in a wall next to Mary].
1 – Jacob tried to shoot at Mary and actually shot at Mary
2 – Jacob tried to shoot at Mary but only thinks he shot at Mary
After answering the prompt “Jacob’s bullet [did/did not] hit Mary”, I would be willing to bet that in the miss condition, a large number of respondents would choose 2.
Hi Simon,
Thanks for your comments! The “shot at” analogy is interesting. But before getting too far down that path, I was hoping to get a better sense of the pragmatic implication which might be causing interference in the study we actually did.
With respect to sentence, “He tried to lie and actually did lie,” almost everyone selected this when the assertion was false, but almost no one selected it when the assertion was true. What pragmatic implication would produce this pattern of results?
It’s also worth bearing in mind that people did not have to interpret “actually did lie” in isolation. They encountered that phrase as a contrast to “only thinks he lied.” For my part, with the contrast clearly in view, I encounter no difficulty in understanding what it means. But I never put too much stock in a finding when N = 1. 🙂
Hi John, thanks. I am not optimistic that we can safely work out pragmatic implications a priori, it would surely be better to test things out and see what happens. But I would suggest that the use of the conjunction implicates that trying to lie and actually lying are not the same thing. So the respondents are in effect being prompted to make a (false) distinction between the two halves of the claim. Another analogy might help:
The ancient Greeks saw a bright star that appeared in the evenings and named it “Hesperus”. They also saw a bright star that appeared in the mornings and named it “Phosphorus”. It turns out that these two “bright stars” were actually the very same planet: Venus. Yesterday evenings, I looked up at the sky and saw the planet Venus. Which of these claims better describes what I saw:
1 – I saw Phosphorus
2 – I did not see Phosphorus
I think people will mostly choose 1 here. But ask a different group which of these claims is the better description:
3 – I saw Hesperus and Phosphorus
4 – I saw Hesperus but not Phosphorus.
And now I’d be willing to bet that a lot of people will choose 4.
Hi Simon,
This really interesting suggestion raises two questions for me. On the one hand, it’s not clear why the false distinction would cause the truth-value manipulation to have such a large effect on people’s responses, effectively pushing people from ceiling (false condition) to floor (true condition). If it’s a false distinction, why do people take the bait when the assertion is true?
On the other hand, if there’s no difference between lying and trying to lie, then what to say about this sort of case? Suppose that there are two grocery stores in town, one on Best Street and one on West Street. Jacob thinks that Mary is at the grocery store on Best Street. The agents ask, “Where is Mary?” Jacob intends to lie by telling them that she is at the grocery store on West Street. However, although Jacob doesn’t realize it, the stress of the situation causes him to flub-mouth. He actually says, “Mary’s at the grocery store on Best Street.”
So he intends to lie by saying “West” and thinks he says “West,” but he actually says “Best.” It definitely seems to me like he tries to lie but fails to lie.
Combine this with the fact that, in general, trying to X does not entail X-ing, and I’m led to think that we aren’t simply giving people a distinction that they antecedently lacked.
John, you’re right of course, there is a distinction between trying to lie and actually lying. I shouldn’t have suggested that there these are identical. (Perhaps your response points to an important disanalogy with my Hesperus/Phosphorus example). What I should have said is just that the distinction is incorrect in the case at hand.
So why are your respondents so strongly tempted to deny this, and say, in the case where he mistakenly says something true, that he tried to lie but only thinks he lied? Perhaps because the story makes salient one distinction: saying something false while intending to say something false/saying something true while intending to say something false, while the descriptions the respondents are asked to apply to the story implicate another distinction: trying to lie and actually lying/trying to lie but only thinking you lied. It is tempting to think the latter distinction maps onto the former (perhaps in part due to norms of conversational relevance: don’t make a distinction unless it’s a relevant one). Since these distinctions are quite unusual and unobvious, it wouldn’t be that surprising if lots of people confuse them. Similarly with the “shot at” case I suggested (which doesn’t depend on identity, as the Hesperus/Phosphorus analogy might).
(By the way, on reflection I think that the above distinction is also a genuine one, though your Best Street story makes obvious that it is slightly different from the actually lying/trying to lie distinction.)
By the way, FWIW I tested the “shot at” case on exactly one conveniently nearby non-philosopher, who without hesitation answered (2). My own view about the ordinary concept of “shot at” remains unshaken!
Hi, Simon. I think it will be hard to test that explanation, but we’ll keep it in mind and perhaps something will occur to us.