On the Ordinary Concept of Lying

Happy New Year, x-philes! 

I have recently been collaborating with Don Fallis (Arizona), who does some excellent work on the concept of 'lying'. Here's the paper, and here's a quick summary of the project and results:

There are, primarily, three popular attempts to define 'lying', all of which claim to be capturing the ordinary concept as used by everyday folk.  The "traditional" definition takes lying to be 'claiming p, where p is false, and where one intends to deceive by claiming p.'  Another definition, put forth and defended by Thomas Carson, defines lying as 'warranting the truth of something that one believes to be false' (where 'warrant' involves implicitly promising or guaranteeing that what one says is true). Finally, Fallis offers his own definition of lying, inspired by Gricean norms of conversation: one lies when one says something that they believe to be false in a context where they believe that there is a norm against saying false things in force.

There are also potential counter-examples, and the debate has largely come down to whether the proposed counter-examples to the various definitions are, in fact, lies.  The "traditional definition" seems to run up against cases of bald-faced lies, where one says something false without the pretext of attempting to deceive the audience (e.g., the student who lies about plagiarizing to avoid expulsion--which, at this school, can only occur when the student has admitted to cheating--even though both she and the teacher know that it's clear the student cheated). Carson's definition, according to Fallis, seems challenged by two different kinds of counter-examples: first, if one adds a proviso (e.g., "but, then, my memory's not so good these days" or "of course, I was pretty drunk at the time") to what would otherwise be a straight-forward lie, the warrant would presumably be defeated and, thus, Carson's definition would rule it out as a lie; second, if one confusedly believes that she is in a serious situation, although she actually is in a non-serious situation (such as a confused politician or a naive Colbert Report guest), and if warrant is a function of the external situation (as Carson seems to believe), then her saying something that she believes to be false would not constitute lying, according to Carson. Thus, both "proviso" lies and "confused" lies would be counter-examples to Carson, given that they in fact constitute lies.

Don and I decided that, since these definitions purport to be about the everyday notion of lying, the best way to settle the debate about bald-faced, proviso, and confused lies was to ask the folk about them.  We had two sets of experiments. In the first experiment, we presented subjects with four short vignettes: one involving a straight-forward lie; one involving a bald-faced lie; one involving a proviso lie; and one involving a straight-forward truth.  For each scenario, subjects were asked to rate on a 7-point Likert scale whether the person lied by saying X.  The results are represented below:

 
Roughly 98% of respondents rated bald-faced lies above the mid-line, and more than 90% rated proviso cases likewise.  Ratings for both bald-faced lies and proviso cases were significantly correlated with ratings for straight-forward lies [(r = .277 (214), p (one-tailed) <.0001) and (r = .171 (213), p (one‑tailed) <.01), respectively]. It seems that, pace Carson and defenders of the traditional definition (e.g., James Mahon), the folk treat both bald-faced lies and proviso cases as genuine instances of lying.

The second experiment utilized a 2x2 design, where we manipulated whether the person saying something they believed to be false (i) was in a serious or non-serious context and (ii) believed he was in a serious or non-serious context.  So, we presented participants with one of four vignettes about a senator running for office who is participating in a series of interviews via satellite: in one condition he believes he is talking to Katie Couric on the Evening News, and is in fact talking to Katie Couric; in another, he believes he's talking to Katie Couric, but is actually talking to Saturday Night Live's Amy Poehler (in character as Katie Couric); in another he believes he's appearing on Saturaday Night Live, and is in fact on SNL; and in the final condition, he believes he is appearing on SNL, but is actually on the Evening News with Katie Couric.  In all of the conditions, the senator says something that he knows to be false about his opponent.  We then asked participants to rate whether the Senator lied:

Interestingly (for reasons unknown but speculated about in the paper), all of the conditions were rated significantly above midline.  More importantly, though, a 2x2 ANOVA revealed that the senator's belief about the situation had a significantly greater impact on ratings than the external facts (F (1, 204) = 40.89, p<.0001).  Participants rated his statement more definitely as a lie when the senator believed he was on the Evening News, regardless of whether he was actually appearing on the Evening News or on SNL.  These results seem to strike a further blow to Carson's definition of lying as warranting, insofar as he cashes out warranting in terms of the external conditions (rather than in terms of one's believing that they are warranting).

Our results suggest that Fallis' definition best captures the ordinary usage of the term 'lie', since it alone can accommodate bald-faced, proviso, and confused lies (while also ruling out presumed non-lies, like performances in plays and jokes).  The paper also discusses the need to disambiguate the semantic content of 'lying' and the moral features.  Feedback is both welcome and appreciated!

Comments

[Lewis Powell]

How was the question to respondents in the first portion phrased? I ask because the provisos you mentioned would themselves be regarded as lies on the Carson account (assuming the person does not have a bad memory and/or was not drunk at the time in question).

[Bryony Pierce]

You say that "Fallis (2009, 34) suggests that you lie if (a) you say p, (b) you believe that p is false, and (c) you believe that the first of these norms of conversation is in force", and that the first norm is that you should not say what you believe to be false. So does this mean that whether people who tell white lies are lying or not depends on whether they think white lies are socially acceptable - whether they believe the first norm is in place? Fallis can't think it's always in place, or acting and sarcasm would be lying, so would he claim that the person who thinks the norm doesn't apply when a white lie seems appropriate is not lying either? I wonder what the folk would say about white lies, and how you view them. Do you accept that lying can be a matter of degree, or do you want a clear distinction between lies and truth? If it can be a matter of degree, should this be made clear in any attempt at a definition?

[Adam]

@Lewis: Good question. We were pretty specific with our wording; since the vignettes were slightly different, each question was slightly different. For instance, in the vignette where Charlie skips out of work to go gambling, the question was 'Did Charlie lie about gambling?'. Given that Charlie added "Of course, I probably wouldn't admit it if I was [gambling]," (which should suffice to defeat warrant) Carson's definition says it does not count as a lie.

@Bryony: Those are some great questions! Really interesting. First, I don't think anyone holds that the conversational norm against saying false things is ALWAYS in effect. Acting, joke-telling, and sarcasm are great examples of that norm being suspended.

Second, I suppose that the point about white lies depends on what we mean by 'white lie'. Are they, by definition, falsehoods told when one thinks that some other (social or conversational) norm turns off the norm of not saying false things? Are they just lies that one believes are morally acceptable (in virtue of some overriding consideration)? I think we usually mean something like the latter when we talk about white lies. One could think that the norm of not saying false things is still in effect, but that some extenuating circumstance makes one think that lying is socially and/or morally acceptable.

As far as whether lies can be a matter of degree, we do say in the paper that it's possible that 'lie' is a prototype or "exemplar" concept, where there is a prototypical picture of a lie, but no set of necessary and jointly sufficient conditions that perfectly captures the content. We think it's pretty likely, actually, that this is the case and that certain lies resemble the prototype more than others. We just happen to think that Fallis' definition best captures the ordinary usage of the term, and the data seem to back that up.

[jonathan weinberg]

(What's with the totally out of control comment spam in the last week? Yikes.)

[Justin Sytsma]

(I wonder if it isn't time for the blog to shift to moderating comments? The downside is that it would be extra work for Thomas; but as there are a lot of contributors, maybe the comment notifications can go to a handful of volunteers, any of which could put the legit comments through and spreading out the work. I would be happy to volunteer for this.)

BTW, looks like an interesting paper, Adam! (Although I haven't had the chance to read the full thing yet.)

[Lewis Powell]

If I follow the post correctly, the three definitions on the table are:

1) Claiming P, where P is false, and one intends to deceive by claiming P.
2) Warranting the truth of something that one believes to be false.
3) Saying something one believes to be false in a context where one believes that there is a norm against saying false things in force.

Has anyone that you know of proposed something like, "claiming what one correctly believes to be false"? This strikes me as an approach similar to (2), but which doesn't tie itself to a particular analysis of claiming (e.g. warranting truth). Intuitively, the bald-faced liar still claims innocence.

With respect to the proviso lie, I would also think that the proviso liar has claimed two things they know to be false, and thus, has not only lied in the first instance (say, about gambling), but also with the proviso (since they do not believe themselves to have a bad memory, etc.).

With respect to the confusion cases (SNL vs. Katie Couric), this analysis would face some difficulty, but I am curious about a) how specific the vignette's were with respect to the false statement (i.e. did the vignette just say "the senator said something false about the opponent's platform" or did it say something more like, "the senator said that the opponent favored universal health care, knowing that the opponent did not support universal health care"). Part of the reason I am curious has to do with whether the credibility of the claim is getting in the way. I'd be curious to see how the numbers shake out between a senator on SNL saying the opponent favors universal health care when they don't vs. a senator on Colbert saying the opponent is a martian (when they aren't).

My thought is that people on SNL may still be seen as making claims, if the claims are within a certain range of credibility. Someone might think, "oh, _I_ know s/he's joking about that, but I'll bet s/he's hoping some of the viewers don't realize s/he's joking". If the numbers were still that high for a vignette in which the senator on colbert is basically spouting science fiction.

[Adam]

Lewis, thanks for the questions.
Working from your last comment first: we do discuss (in the paper) a number of potential issues surrounding the confused politician cases, including the possibility that subjects might have reasoned similarly to your proposal (i.e., that they might have thought a political message was still being communicated under the guise of a joke). We tried to make the claims severe enough to be taken as parody in a non-serious context, but realistic enough to be taken as campaigning in the serious context. The claims made were (a) that his opponent wants to invade Spain in order to fight terrorism, and (b) that his opponent wants to kill newborn babies in order to promote stem cell research. I'm curious: do you think those are too extreme, or not extreme enough?

W/r/t what you said about the provisos, the provisional addendum wasn't always a false claim. Indeed, in many of the scenarios, the proviso was a sort of blatant honesty: "Of course, I probably wouldn't admit it if I had, but..." or "...not that I'd tell you if I did." And in the others, it's not clear that the provisional addendum isn't true (the vignette does not preclude, for example, the speaker having a bad memory). Moreover, the questions were very specific about what statement was under consideration (e.g., "Did Jamie lie to Chris about who broke the trophy?")

Now, as far as your proposed definition goes, I'm not familiar with all the literature (that's Don's expertise), but I suspect that it's not going to be able to accommodate joke-telling and sarcasm (and maybe not acting, either), since both of those are instances of claiming actual falsehoods, but neither of which is typically considered a lie. Additionally, one might think it's possible to lie, even if one is unwittingly telling the truth. I don't have data on that, but the major result of the second study was that participants put a greater weight on the speaker's intentional states than on the external states of affairs. At the least, that's prima facie support for unwitting truths potentially counting as lies.

[Lewis Powell]

Adam,
thanks for your replies.
As to the campaign lies, I'd say the invading spain example is _less_ extreme than I was imagining, but the infanticide example is in the neighborhood.

As to the provisos: good points,

As to the definition I proposed: I am pretty sure it can capture straightforward joke-telling. If I say "a priest and a rabbi walked into a bar", I don't think I've claimed that a priest and a rabbi have walked into a bar. For sarcasm, I am less sure, but again, if Smith is usually a source of interpersonal conflict at the workplace, and somehow gets put in charge of conflict resolution at the workplace, a coworker uttering "yeah, this will end well" does not seem to have claimed that it will end well. In Grice's terminology, the co-worker would have "made as if to say" that things will end well, in order to communicate the opposite. I'd also think the definition captures acting (for similar reasons).

[Lewis Powell]

It occurs to me that my proposed definition may not be incompatible with Fallis's, depending on the analysis of claiming.

If claiming P is uttering a sentence that means P while believing that there is a norm to say true things in effect in your present context, then my definition is compatible with Fallis's (except with respect to the external condition whereby mine requires that the person say a falsehood). I personally tend to think of lying as involving the utterance of a falsehood, but I didn't mean for that to be the crucial difference between the proposal I gave and Fallis's. My thought was that some of the issues that are coming up about lying are broader issues about claiming or asserting (i.e. under what conditions does uttering "S" produce an assertion of P), and that various proposals can be understood as completions of the schema:
"In uttering 'S' one lies just in case i) the utterance produced an assertion of P and [...]"

Where "..." could be filled in by "P is believed to be false" or "the intention of the utterance is to deceive the audience" etc.

If we allow that lying can be factored into an assertion/claim component and some further condition(s), it looks like Fallis and Carson diverge on the analysis of asserting or claiming, not on the further condition. The traditional conception differs from both of theirs on the further conditions, both in that it requires the intent to deceive (rather than merely requiring the belief that the content is false) and in that it requires the falsity of the thing asserted (rather than not imposing an external condition).