ArticlePDF Available


A new definition of lying is gaining traction, according to which you lie only if you say what you know to be false. Drawing inspiration from "New Evil Demon" scenarios, I present a battery of counterexamples against this "Knowledge Account" of lying. Along the way, I comment upon the methodology of conceptual analysis, the moral implications of the Knowledge Account, and its ties with knowledge-first epistemology.
Lying: Knowledge or Belief?
Neri Marsili*
Forthcoming in Philosophical Studies
A new definition of lying is gaining traction, according to which you lie only if you say what
you know to be false. Drawing inspiration from “New Evil Demon” scenarios, I present a battery of
counterexamples against this “Knowledge Account” of lying. Along the way, I comment upon the
methodology of conceptual analysis, the moral implications of the Knowledge Account, and its ties with
knowledge-first epistemology.
Definition of Lying; Insincerity; Knowledge-first Epistemology; Ethics of
Lying; Metaphilosophy; Conceptual Analysis
1. Lying and knowledge
He who cannot lie, doth not know what truth is
Also Sprach Zarathustra
Suppose I tell you something that I believe to be false: for instance, that it is raining in Athens. If I
believe (say, with good reasons) that it is not raining in Athens, but I don’t in fact know that it is not
raining there, am I lying? More generally, is lying a matter of asserting what you believe to be false, or
of asserting what you know to be false?
The traditional view is that the notion of lying is tied to belief, rather than knowledge.
Philosophers typically accept that (A) and (B) are necessary
conditions for lying:
S lies only if:
(A) S asserts that p
(B) S believes that ¬p
Recently, however, it has been argued that (B) is too lenient. The suggestion is that lying should
rather be tied to the notion of knowledge. To return to our initial example, I would be lying only if I
knew that it’s false that it’s raining in Athens; disbelieving it is not enough for my utterance to be a lie.
If this is correct, (B) should be replaced with (K):
(K) S knows that ¬p
Call the view that lying requires (K) instead of (B) the Knowledge Account of Lying. This view is only
recently gaining traction, but it does have its precursors. To my knowledge, its earliest formulation
is in Coignet's Politique discourses upon Truth and Lying (1586), where lying is defined as a communicative
act whereby “one speaketh of things vncertain, contrarie to that which one knoweth”
. Note,
however, that Coignet only requires that the speaker say something that contradicts what they know,
* Alma Mater Studiorum Università di Bologna. Contact:
There is disagreement about whether sufficiency requires that S intends to deceive the audience (Lackey 2013; Meibauer
2014; Mahon 2015; Fallis 2018; Dynel 2018; Stokke 2018; Krstić 2019; Wiegmann and Meibauer 2019, Marsili 2021, §2.3).
This issue can be safely set aside for our purposes. For an approach to defining lying that relies on identifying prototype elements
instead of necessary and sufficient conditions, see Coleman & Kay (1981).
Some of Sartre’s remarks about bad faith point in a si milar direction: “The essence of the lie implies in fact that the liar
actually is in complete possession of the truth which he is hiding. A man does not lie about what he is ignorant of” (Sartre
1956, 48).
Forthcoming in Philosophical Studies
which is weaker than (K). For a formulation closer to (K) we have to wait until the XXI century.
Saul (2000) is probably the first to claim, in passing
, that ‘lying requires knowingly speaking falsely’.
Caminada (2009) endorses (K) more explicitly; like Saul, however, he doesn’t motivate his preference
for (K) over (B). Its only in the last decade that the Knowledge Account has been explicitly
motivated and defended, in the works of Turri (2016, 3436), Benton (2017), and Holguín (2019).
The question of whether lying requires knowledge or belief may seem philosophically marginal,
but there is more at stake than the analysis of a concept, like lying, that is central to human
communication. First, there are ethical implications. If (against traditional views) lying requires
knowledge rather than belief, we would have to significantly rethink how we understand the moral
blameworthiness of lying: whether one is responsible for having lied would depend not only on what
one believes, but also on factors that are not (or that are only partially) under the speaker’s control,
such as whether the proposition is actually false, whether the belief is Gettiered, and so forth. This
isn’t a straightforward take on what makes lying blameworthy, and we will see that assessing the
merits of this view requires some sophisticated reflection on how intentional action is assessed more
Second, if the Knowledge Account turned out to be correct, this result would speak in favour
of the knowledge-first paradigm in epistemology: the idea, gaining traction over the last two decades,
that knowledge is a more fundamental concept than belief – an ‘unexplained explainer’ that can help
us understand other key concepts that are central to epistemological theorising (Williamson 2000;
McGlynn 2014), including lying. If the notion of lying is indeed best understood in terms of
knowledge rather than belief, we would have new proof of the explanatory power of knowledge-
first epistemology; if not, we may see this as a novel reason to question its ability to revolutionise
our way of thinking about basic philosophical concepts.
I will come back to both points later in the paper. In the next section, I present and assesses the
positive case for preferring (K) over (B). Section 3 shows that the Knowledge Account is subject to
several counterexamples that are avoided by traditional definitions. Section 4 explores the moral
implications of the Knowledge Account in detail. Section 5 introduces a last set of counterexamples,
and section 6 considers the viability of defending the Knowledge Account by appealing to error
theories. Section 7 concludes.
2. Reasons for preferring knowledge over belief
Why should we think that (K) is preferable to (B)? Benton, Holguín, and Turri defend the
Knowledge Account by appealing to what we may call coherence arguments. Broadly, the idea is that we
should endorse the Knowledge Account because it coheres with philosophical views that are
independently supported by sound arguments and compelling empirical evidence.
First, it is noted that the Knowledge Account sits well with knowledge-first epistemology. A key
contention of knowledge-first epistemology is that several concepts that are central to
epistemological theorising are best explained in terms of knowledge, rather than (e.g.) belief
(Williamson 2000, ch.1-2; McGlynn 2014, 16). If one is sympathetic to knowledge-first epistemology,
one has a motivation to prefer (K) over (B). Second, the Knowledge Account has on its side its
“elegant simplicity and pleasing symmetry” (Turri 2016, 34). Third, and perhaps most importantly,
it is reinforced, indirectly, by the philosophical and empirical case in favour of the knowledge-norm of
Let’s focus on this last point, since it needs unpacking. The idea is that the Knowledge Account
is a natural way to define lying for those who accept the knowledge-norm of assertion: the view that
you are epistemically entitled to assert that p only if you know that p. As Holguín (2019, 5352) puts it,if we
think there are knowledge-theoretic norms of assertion, we should expect the concept of lying to be
intimately connected with the concept of knowing”.
Here’s a more explicit formulation of the idea (following Turri 2016, 34). Suppose that
knowledge is the norm of assertion: you are entitled to assert that p only if you know that p.
Whenever you assert a proposition that you don’t know, you violate the knowledge-rule. Suppose,
further, that the term lying denotes a special infraction of the norm of assertion: we use it to
designate “cheating” at asserting. Arguably, to ‘cheat’ one must knowingly (that is, intentionally and
Saul (2012) has since moved on to defend a traditional view, p referring (B) over (K).
For a critical overview of the philosophical case for the knowledge-norm, McGlynn (2014, chap. 5); for the empirical
case, Turri (2017).
successfully) break a rule. It follows that you lie when you knowingly break the rule of assertion
when you assert what you know to be false.
However, this argument only seems to establish the desired connection between (K) and cheating
at assertion’. If we accept its premises, we can only conclude that you lie when you know that you don’t
know that p. This isn’t enough to prefer (K) over (B), because this requirement is weaker than both
(K) and (B) it can be met by a speaker who sincerely believes that p and doesn’t know that p is
. Turri’s argument for deriving (K) from the knowledge-norm of assertion falls short of its
intended goal
Suppose, however, that we find this set of coherence arguments persuasive. Even so, there are
two important limitations that apply to them. First, coherence arguments offer at best conditional
reasons to accept the Knowledge Account. This is a problem because both the knowledge-norm of
assertion and the knowledge-first paradigm have been subject to substantial criticisms
, and neither
can be considered a consensus view. If these criticisms are sound, and knowledge isn’t the norm of
assertion (nor is it more fundamental than belief), we lack a positive case to prefer (K) over (B)
because the antecedent of our conditional argument is false. The result is that the positive case in
favour of (K) is pretty weak: only if a certain set of views is correct, we have a positive reason to
prefer (K) over (B)and since the correctness of these views is disputed, it is equally disputed that
we enjoy this positive reason to prefer (K) over (B).
Second, even if the antecedent of the conditional argument is true (and we have a positive reason
to accept the Knowledge Account), the problem remains that coherence arguments alone cannot
provide us with conclusive reasons to accept the Knowledge Account. This is because philosophers
tend to agree that intensional accuracy (i.e. whether a definition systematically tracks our intuitions) is
the primary desideratum for a definition of lying (Fallis 2009, §1; cf. Gupta 2015, §1.4), and that this
desideratum must be met for a definition to be informative in our moral theorising about lying
(Marsili 2019). Clearly, we cannot determine that a definition is intensionally accurate on the basis
of coherence arguments alone. If it turns out that replacing (B) with (K) leads to predictions that
systematically clash with our semantic and moral intuitions about lying, insisting that (K) coheres
with existing philosophical views will do little to prove that the Knowledge Account is a correct
definition of lying (rather than, say, a useful theoretical construct).
Summing up, the positive case for the Knowledge account is somewhat weak. We cannot easily
derive (K) from the knowledge-norm of assertion. And coherence arguments can only provide
conditional, non-conclusive reasons to accept the Knowledge Account. Of course, this only means that
we have little reason to endorse the view, not that we should reject it. But we will now see that there is
also a solid negative case against the Knowledge Account.
3. The New Evil Liar
Here’s a first way to challenge the idea that we can only lie about what we know to be false. Consider
the following scenario, in the spirit of New Evil Demon scenarios (Littlejohn 2009):
Pinocchio saw that Lucignolo stole an apple from his classmate Berlicche. Based on this
experience, he forms the correct belief that Lucignolo stole Berlicche’s apple. When
Pinocchio notices that Berlicche is looking for his apple, he tells him:
(1) Eugenio stole the apple from your desk
Pinocchio has a twin, Malocchio, who is the unfortunate inhabitant of an evil demon world.
Based on experiences indistinguishable from those of Pinocchio, Malocchio forms the belief
For example, suppose that I believe p on a hunch, for no good reason. I know that I don’t know that p. But if I assert p, I
would meet neither (K) nor (B).
Benton (2018) proposes a similar derivation, but acknowledges the argument’s limitations. Holguín (2019) derives (K)
from the knowledge norm and the stronger premise that one lies when “one asserts the negation of what one ought to have
asserted”. This gets us to (K), but at the price of quasi-circularity: Holguín’s extra premise (for which no argument is provided)
is just (K) in disguise.
See, e.g., Pagin (2016) and Marsili (2018b) for criticisms of the former, and McGlynn (2014) and Gerken ( 2017) for
criticisms of the latter. A criticism of the empirical case in favour of the norm of assertion is in Marsili and Wiegmann (2021).
Forthcoming in Philosophical Studies
that Lucignolo stole an apple from Berlicche. Just like Pinocchio, Malocchio tells Berlicche
that (1).
Intuitively, both Pinocchio and Malocchio have lied to their respective classmates. However, while
Pinocchio has asserted something that he knows to be false, Malocchio cannot know that what he
said is false, because he inhabits an evil demon world. According to the Knowledge Account,
Pinocchio lied, but Malocchio didn’t.
This seems intuitively wrong
. Pinocchio’s and Malocchio’s actions were motivated by the same
evidence and doxastic states, and by the same malicious intentions: intuitively, if one has lied, also
the other one has. The Knowledge Account is committed to denying this. Since the two scenarios
are identical from the speaker’s perspective, the Knowledge Account also entails that whether a
speaker is lying may depend on factors that are beyond the speaker’s control. This is another
controversial commitment, which has problematic moral implications. If (as the Knowledge
Account has it) only Pinocchio lied, his behaviour is morally criticisable in a way in which
Malocchio’s behaviour is not: both would be guilty of attempted deception, but only Pinocchio would
be guilty of lying. Intuitively, however, Pinocchio and Malocchio are equally blameworthy: after all,
they acted in the same way, with the same intentions and consequences, and for the same reasons
This places an additional burden on the Knowledge Account: its proponents need to offer a plausible
story for introducing a morally loaded distinction where there seems to be none.
Admittedly, New Evil Demon scenarios are no everyday situations, and as such it may be
maintained that they have limited evidential weight. In response, we could consider a more realistic
NED scenario:
Pinocchio owes 15 liras to Lucignolo, who asks for his money back. Pinocchio knows that
he put 20 liras in a hidden pocket of his schoolbag just before leaving home. Nonetheless,
he tells Lucignolo:
(2) I don’t have 15 liras on me right now
Pinocchio has a twin, Bucocchio, who lives in a different town. Like Pinocchio, Bucocchio
owes Lucignolo’s twin 15 liras, has 20 liras hidden in his pocket, and has told Lucignolo’s
twin that (2).
But Bucocchio is unaware that his pocket has developed a hole since he last used it. Some of
Bucocchio’s money has fallen out of his pocket since he left home.
Holguín (2019) discusses a similar case, and attempts to explain away this prediction of the Knowledge Account. I will
discuss his arguments in sections 4 and 6. A referee suggests that, since the internal processes underlying Malocchio’s belief
are isomorphic to Pinocchio’s, it may be argued that both meet the standard for knowing and, thus, that both are lying in uttering (1).
In reply, even if we conceded that Malocchio knows that (1) is false, we could not infer from this that he is lying: this inference
would treat (K) as a premise, but our aim here is precisely to est ablish whether (K) holds. Furthermore, this strategy is unlikely
to generalise. I will soon discuss NED scenarios where the twin lacks different requirements for knowledge (safety, truth,
justification, confidence, etc.). To argue that the twin can meet the standard for knowing in each case would force us to revise our
concept of knowledge beyond recognition, and (K) would end up being a very different thesis (one much closer to (B)).
Additionally, and crucially, in §4 I will consider a case that does not involve internally isomorphic twins. For more on
revisionist replies, cf. §6.
Can we avoid this prediction simply by assuming that lying is just as morally bad as misleading? Endorsing this view
may help narrow the gap between the predictions of the Knowledge Account and our moral intuitions and practices, but it
won’t fill the gap altogether. Even authors who defend the thesis that misleading is just as bad as lying acknowledge (i) that
lying and misleading are different kinds of infractions, and (ii) that, as a matter of fact, we associate different moral stigmas
and different social sanctions to each (cf. Saul 2012, esp. §4.4). The Knowledge Account’s verdict that Pinocchio and
Malocchio have committed di fferent kinds of infraction is therefore problematic regardless of one’s take about the moral
difference between lying and misleading, since it clashes with both (i) and (ii). For more on the moral implications of (K), see
Also in this scenario, both Pinocchio and Bucocchio intuitively lied. However, Bucocchio doesn’t
know that what he said is false. The Knowledge Account rules that Pinocchio has lied and Bucocchio
hasn’t. But this doesn’t seem right: after all, this scenario is structurally similar to NEW EVIL LIAR,
except that it involves local deception (Bucocchio is mistaken about his finances), rather than
universal deception (Malocchio knows nothing about the external world). Furthermore, this
classification licenses morally loaded judgments that seem equally doubtful: while Pinocchio would
be guilty of lying, Bucocchio could at most be criticized for attempting to deceive.
Note that the scenario does not specify whether Bucocchio actually has fifteen liras. If more
than 5 liras fell out of Bucocchio’s pocket (TRUE-VERSION), Bucocchio doesn’t know that (2) is
false because (2) is true. If less than 5 liras fell out of Bucocchio’s pocket (GETTIER-VERSION),
Bucocchio doesn’t know that (2) is true because his belief is unsafe and based on misleading evidence.
I left both options open because not all readers will find the TRUE-VERSION to be a
straightforward case of lying. While most philosophers
and laypeople
would agree that
Bucocchio is lying even if it turns out that more than 5 liras fell out of his pocket, a few readers may
not share this intuition. But this is no reason to find the GETTIER-VERSION any less convincing. In
the GETTIER-VERSION there is even less difference between what Pinocchio and Bucocchio have
done (both have said something false), making it even less plausible to claim that one is lying while
the other isn’t
. The NEW EVIL HOLE scenario is effective because it offers a counterexample for
every taste: most people will find both versions convincing, and the few who are unpersuaded by
TRUE-VERSION should still be persuaded by its GETTIER-VERSION
4. Blameworthiness and internal requirements for knowledge
I have argued the Knowledge Account makes predictions that are hard to reconcile with (i) our
intuitions about what lying is, and (ii) our intuitions about the moral difference between each pair of
examples. Here I would like to consider whether something can be said in response to the second
The accusation is that the Knowledge Account introduces a morally loaded distinction where
there seems to be none. For instance, many will intuit that whether you are criticisable for having
lied should not depend on whether your belief in what you said is Gettiered or not. Controversially,
the Knowledge Account rules otherwise. Can something be said to explain away this prediction?
Here’s a way to respond on behalf of the Knowledge Account: in both the “good cases” (where
the speaker has knowledge) and the “bad cases (where the speaker lacks it) of every scenario
considered so far, the speaker believes that they are lying that is, they believe that they know that what
they said is false. For instance, in NEW EVIL LIAR (where Malocchio’s belief is Gettiered, but not
Pinocchio’s), both Pinocchio and Malocchio believe that they know that their assertion is false. If we
have the intuition that Pinocchio and Malocchio are equally blameworthy, it may be suggested, it is
not because they both lied, but rather because they both believe that they lied.
Holguín (2019, 5366) has hinted at the possibility of a reply along these lines. He mentions that
the moral implications of the Knowledge Account may be reconciled with our intuitions by
appealing to a principle like (P):
(P) If as a result of X-ing A is (or would be) subject to normative scrutiny, then for any other
person B that is in the relevant respects an internal duplicate of A, B is subject to some amount of
normative scrutiny too
Augustine (DM), Aquinas (ST), Kant (1797), Leonard (1959), Isenberg (1964), Mannison (1969), Chisholm and Feehan
(1977), Kupfer (1982), Adler (1997), Williams (2002), Mahon (2015), Horn (2017), Wiegmann (forthcoming). For the
opposite view, Sartre (1956); for intermediate views, Siegler (1966), Carson (2010), and Saul (2012).
Turri and Turri (2015) found mixed evidence about people’s intuitions: participants would respond in radically
different ways, depending on how the question was asked. A subsequent study by Wiegmann, Samland, and Waldmann (2016)
has shown that people consistently judge true lies to be genuine lies, and that the mixed findings by Turri and Turri were the
by-product of a flawed questioning method (cf. Marsili 2016, 299, Wiegmann, Rutschmann, and Willemsen 2017, §3,
Wiegmann and Meibauer 2019). In light of these criticisms and further empiri cal research, Turri and Turri (2019) retreated
to the weaker (and much less controversial, cf. Coleman and Kay 1981) claim that only prototypical lying requires falsity.
We may even construct the scenario so that no liras fell off the pocket: Bucocchio’s belief would be lucky and unsafe
anyway, preventing it from being knowledge (Sosa 1999; Pritchard 2005).
Of course, those who intuit that Bucocchio is not lying in the TRUE VERSION will be dissatisfied with both (K) and
(B), since (K) makes incorrect predictions in the GETTIER VERSION, and (B) in the TRUE VERSION. To accommodate the
intuition that lies must be false while avoiding the pitfalls of (K), a further condition (e.g. T: the proposition is false) could
be added to the definition.
Holguín (2019, 5366). Italics are mine.
Forthcoming in Philosophical Studies
If (P) is correct, we would have an alternative explanation for our intuition that Pinocchio and
Malocchio are equally blameworthy: the reason is not that they both lied, but rather that their actions
were the result of internally identical processes: both intended to bring about the same consequences,
and both thought that they were committing the same morally reprehensible act.
This explanation may seem promising, but there are two important difficulties that prevent it
from being fully persuasive. The first is that (P) is too weak. It only tells us that an internal duplicate
is subject to some normative scrutiny, not that internal twins are subject to equal normative scrutiny.
This doesn’t help much: recall that the intuition that we need to explain away is that Pinocchio and
Malocchio seem to have committed the same kind of infraction, and therefore that they are equally
morally blameworthy, and liable to the same sort of criticism.
To address this worry, we could amend (P), and replace “some” with “the same”. This modified
version of (P) call it (P*) would correctly predict that the twins in our examples are equally
blameworthy. However, (P*) is a much more controversial principle than (P), and as such it stands
in need of proper justification. To illustrate its contentiousness with an example: according to (P*),
attempted murder should be subject to the same moral scrutiny as successful murder. While there is
a sense in which both attempted and actual murder reflect equally badly on the moral character of an
agent, it is at least disputable that murder and attempted murder are equally morally reprehensible,
or liable to the same sort of criticism
. While (P) is too weak, there is a risk that (P*) may be too
Suppose that we accept (P*) despite these worries. We would then face a second, more pressing
problem: namely, that counterexamples to the Knowledge Account need not involve internal
duplicates. To illustrate, consider the following scenario:
Grandpa Joe recently died. Pinocchio, who’s very superstitious, organises a seance with
a medium to try to communicate with Grandpa Joe one last time. During the seance, its
revealed that Grandpa Joe left all his money to a woman. On this basis, Pinocchio forms
a firm belief that Grandpa Joe left all his money to Auntie Sue.
Pinocchio was in fact meant to see a notary today – to discuss, amongst other things,
Grandpa Joe’s will. Believing that Pinocchio has already met with the notary, Auntie Sue
calls Pinocchio, asking if there are any news about the will. Pinocchio is mad at Auntie
Sue. Out of pure spite, he replies:
(3) It turns out that Grandpa Joe left all his money to charity
Auntocchio inhabits a twin world where everything holds the same. The only difference
is that Auntie Sue calls him after Auntocchio saw the notary, and that the notary has
confirmed that Grandpa Joe left everything to Auntie Sue.
Both Pinocchio and Auntocchio have intuitively lied to Auntie Sue, and they are both criticisable
for having lied. However, while Auntocchio knows that what he said is false, Pinocchio only believes
that what he said is false he cannot know it, for his belief is based on unsafe evidence
. The
Knowledge Account incorrectly predicts that only Auntocchio lied, and therefore that only
Auntocchio is blameworthy for having lied.
The example is interesting for two reasons. First, because it provides a novel kind of
counterexample to the Knowledge Account. In this scenario, Pinocchio’s belief is not merely unlucky.
It is based on blatantly poor evidence. Furthermore, Pinocchio and Auntocchio are not internal
duplicates: their perspective on the external world (and the evidence available to them) differs
significantly. While Auntocchio has access to good, safe reasons (the notary’s testimony) to believe that
what he said is false, Pinocchio has only access to poor, unsafe reasons (the séance) in support of the
same belief.
On the moral and legal difference between successful and attempted crimes, see e.g. Davis (1986). For an overview
on moral luck more generally, see Nelkin (2021). Note that I am not here denying that (P*) is defensible; I am merely pointing
out that (P*) is a costly commitment, since (P*) takes a rather strong stance on philosophically controversial issues.
I am assuming that both Pinocchio and Auntocchio inhabit a world (like ours) where it is impossible to communicate
with dead people: if Pinocchio is right about Grandpa’s will, this is due to luck, not to sound reasoning.
Second, given that Auntocchio and Pinocchio are not internal duplicates, we cannot appeal to
(P) or (P*) to explain away the intuition that Pinocchio and Auntocchio are guilty of the same kind
of infraction. (P) and (P*) have simply nothing to say about cases that don’t involve internal
duplicates: our moral intuitions about this scenario cannot be reconciled with the Knowledge
Account by appealing to these principles.
Recapitulating, the worry outlined at the beginning of this section seems hard to address. The
Knowledge Account does face an explanatory burden that the Belief Account does not face: only
the former view needs to be complemented with some principle that explains away our moral
intuitions about the cases in which the speaker satisfies (B) but not (K). A principle like (P*) isn’t
likely to help: beyond being difficult to defend, it is useless for scenarios, like EVIL AUNTIE, that do
not rely on internal duplicates. It seems difficult, then, to reconcile the Knowledge Account with
our moral intuitions
. Judging by moral implications alone, the Belief Account fares better than its
knowledge-based counterpart.
5. Graded-belief lies
Knowledge is incompatible with uncertainty: you know a proposition only if you outright believe that
. But is outright belief required for lying? It has been argued that this requirement
would be too strong (Carson 2006, 2010; Whyte 2013; Marsili 2014, 2018a, 2019; Krauss 2017; Trpin,
Dobrosovestnova, and Sebastian 2020; cf. Isenberg 1964; Benton 2018). If (K) requires outright
belief, it incorrectly rules out graded-belief lies. Consider the following scenario:
Currocchio is having friends for dinner. A guest asks him if the curry contains nutmeg, explaining
that she has a mild intolerance. Currocchio is quite confident that he did put nutmeg in the curry,
although he isn’t sure: he cooked the curry the previous day, and added lots of random spices.
Since he can’t be bothered preparing something new, he simply replies:
(4) I didn’t put any nutmeg in the curry
Currocchio plainly stated something that he takes to be likely false. In doing so, intuitively he lied.
There is consensus that this is the right verdict: while philosophers disagree about which definition
best handles graded-belief lies like (4), there is agreement that graded-belief lies are intuitively lies
. A
definition that rules that Currocchio has not lied is perhaps defensible, but in need of a strong
independent argument to support its verdict. Once again, (K) is in a weak position to explain our
6. Explaining counterexamples away
We have seen that the Knowledge Account clashes, sometimes even sharply, with our pre-theoretical
intuitions about what lying is. If so, how can its proponents maintain that (K) is preferable to (B),
the traditional view? Among the defenders of (K), only Holguín (2019) has attempted to provide an
Holguín’s strategy is to question the reliability of our intuitions about each example. For instance,
in response to counterexamples based on graded-belief lies, he suggests (i) that lying and bullshitting
Earlier I considered the idea that believing that you committed a moral infraction is sufficient for being blameworthy
for that action. One may wonder whether it is rather this principle that can explain why the twins in each scenario are equally
blameworthy. Leaving aside the fact that this alternative principle is less plausible than (P*), I shall note that it is easy to
amend EVIL AUNTIE so that this requirement is not satisfied: we simply need to specify ex hypothesi that Pinocchio doesn’t
believe that he knows that (3) is false. For instance, Pinocchio may be aware that mediums are unreliable sources, but form
his belief (perhaps irrationally) despite being aware of this. More generally, this alternative principle will not work whenever
the speaker holds a belief despite being aware that that belief is based on poor or unreliable evidence.
This is a requirement that knowledge-firsters accept. See, for instance, Williamson (2000, §4).
For a detailed defence, see Krauss (2017) and Marsili (2014, 2018a, 2019), who provide competing explanations for
this datum. See Holguín (2019, 5361-4) and §6 of this paper for an argument on whether we should doubt the intuition that
graded-belief lies are genuine lies.
(K) is not alone: since (B) too requires outright belief, also (B) incorrectly predicts that Currocchio isn’t lying. Luckily,
there are alternative views that avoid this unpalatable prediction, on which I will come back in the next section.
Forthcoming in Philosophical Studies
are mutually exclusive categories, and (ii) that “S bullshits iff S knows that S does not know whether
p”. These stipulations entail that graded-belief lies are not genuine lies, but rather bullshit, so that our
intuitions are mistaken whenever we classify a statement like (4) as a lie. If our intuitions about
graded-belief lies like (4) are mistaken, such ‘counterexamples’ pose no real challenge to (K). Similar
error theories are introduced to argue that our intuitions about Gettiered lies (like NEW EVIL LIAR)
and ‘true lies’ (like the TRUE-VERSION of NEW EVIL HOLE) are systematically unreliable.
Generalising, the idea is that (K) can be defended by introducing error-theories that explain away
our intuitions about each counterexample. Call a response along these lines an ‘error-theoretic
Considering the merits of each error theory introduced by Holguín in defense of (K) would
require a long digression, but I worry that such a detour would be both unnecessary and ineffective.
It would be unnecessary because this paper has already introduced counterexamples to (K), such as
the EVIL AUNTIE case, that evade Holguín’s defense
. Even if Holguín’s error-theories were all
successful in explaining away their target counterexamples, we would still have that (K) makes
incorrect predictions about all those cases in which the speaker meets (B), but has undisputably poor
evidence in support of their belief (like in EVIL AUNTIE).
Discussing Holguín’s proposed cocktail of error theories would also be an ineffective way to
attack the Knowledge Account, because it would not rule out the possibility that (K) could be
defended by appeal to some other combination of error theories. If we could prove that error-
theoretic responses are in principle improper, by contrast, we would have a much stronger argument
against the Knowledge Account. In the rest of this section, I will attempt to offer a general reply
along these lines.
Let me start by noting an apparent tension between the methodology of conceptual analysis and
the idea that we can explain away counterexamples by appealing to error theories about our
intuitions. We have seen (§2) that the main goal of authors who attempt to define lying is to develop
an analysis of this concept, which requires tracking ordinary language intuitions about whether a
given utterance is a lie. Within this kind of inquiry, intuitions are the raw data against which
competing definitions are tested
. Postulating error theories about our intuitions amounts to
excluding some of the data (typically, and problematically, data that would falsify one’s favourite
view). If postulating error theories about our intuitions is acceptable at all when engaging in
conceptual analysis (as opposed to conceptual engineering), then, it would seem that it is when it is
done (i) parsimoniously, (ii) with sound motivations, and (iii) in absence of more economic
alternatives. The envisaged strategy for defending (K) doesn’t meet any of these conditions.
If we are to rescue (K), error-theories cannot be postulated sparingly. This point is illustrated by
Holguín’s (2019) own defence, which has to
harvest, one after another, several revisionist theories
about our intuitions about knowledge and lying (e.g. about the difference between lies and true lies,
the difference between lying and bullshit, the possibility of unjustified knowledge). Unsurprisingly, being
generous with such revisionary assumptions generates difficulties. First, by simple probability
calculus, each ancillary assumption automatically decreases the probability that the conjunction of
views is correct. Second, each ancillary assumption brings us closer to providing a revisionary
account of lying, and thus farther from our aim (a descriptive analysis of the concept). Third, the
motivation to endorse any specific combination of revisionary assumptions (rather than another)
cannot be that that specific combination squares our intuitions with the Knowledge Account, on
pains of circularity. But it isn’t clear what other reasons we might have to adopt such a specific
collection of revisionary views. The charge of ad-hocness isn’t inescapable here, but it lurks around
the corner for the defender of the Knowledge Account – and existing defences, like Holguín’s, have
done little to counter it.
Arguably, also EVIL CURRY is a counterexample for which no error theory has been provided yet. Holguín only
discusses graded belief in relation to lottery cases, where the speaker’s belief is formed on the basis of statistical evidence.
EVIL CURRY relies on a different kind of uncertainty (compromised memories), and a plausible case can be made that
Holguín’s argument is less effective when applied to this sort of case. For a further counterexample to (K) for which Holguín
provides no error theory, cf. Marsili’s (2016) empirical study on believed-true lying promises.
This is something on which there is solid agreement within the literature, and this very assumption is what motivates
the empirical research that has been conducted about laypeople’s concept of lying. For discussion, see Carson (2006, 285;
2010, 33-7), Fallis (2009, 32), Arico & Fallis (2013, §3), Weissman & Terkourafi (2018, §4), Marsili (2019, §2), Reins &
Wiegmann (2021, §2.3).
The emphasis is on ‘h as to’. (K) imposes several requirements on lying: that the lie be false, that the speaker’s belief be
safe, justified, not graded, etc. Since each requirement is challenged by independent counterexamples, equally independent error-
theories have to be introduced to explain them all away.
Perhaps these criticisms would be less pressing if we lacked more intuitive accounts of what lying
is. But (and this is the crucial point) this isn’t the case. There are definitions that deliver the right
classifications at a much lower costthat is, without invoking an array of exotic error theories. The
Outright Belief account (B) avoids all the counterexamples discussed in section 3 and 4 (NEW EVIL
LIAR, NEW EVIL HOLE, and EVIL AUNTIE). And if one is moved by counterexamples based on
graded-belief lies (EVIL CURRY), more fine-grained accounts are available, such as Carson’s (2006,
2010) Weak Insincerity Condition, or Marsili’s (2014, 2019) Comparative Insincerity Condition
(WIC) The speaker does not believe the proposition to be true
(CIC) The speaker takes herself to be more confident in the falsity of p than in its truth
These accounts can deal with all the cases discussed so far, without postulating any error theories
about our intuitions. They avoid counterexamples inspired by ‘New Evil Demonscenarios, because
they do not require knowledge. And they accommodate graded-belief lies, because they do not
require outright belief
In short, the contention here is that any putative error-theoretic defence of the Knowledge
Account will not meet any of the above identified conditions (i-iii). If the aim is to provide an analysis
of the concept of lying (as opposed to a revisionist account), views based on belief (like B) or
credence (like WIC or CIC) should be preferred, since they track our intuitions far more reliably
than the Knowledge Account.
7. Belief rather than knowledge
The Knowledge Account of lying is an appealing view. It sits well with influential philosophical
theories, such as the knowledge-rule of assertion and the knowledge-first paradigm in epistemology,
and has on its side its “elegant simplicity and pleasing symmetry”. But a good definition must also
be correct that is, intensionally accurate. We have seen that Knowledge Account fails in this respect:
it incorrectly excludes some straightforward cases of lying, and has counterintuitive moral
implications. While it may be useful as an artificial theoretical construct, it isn’t helpful as a
descriptive analysis of what lying is.
We can finally answer our initial question: is lying a matter of asserting what you believe or know
to be false? Perhaps surprisingly, we are better off sticking with traditional views: lying is a matter of
saying what you disbelieve, rather than what you know to be false. Similarly, the distinctive moral
blameworthiness associated with lying is intuitively a matter of internal doxastic states, and seems to
be independent of any further, non-doxastic conditions for knowledge (such as whether the belief
is correct, whether the reasons supporting it are safe, reliable, etc.).
These results, in turn, put some pressure on the ambitions of knowledge-first epistemology. The
traditional, belief-first analyses of the concept of lying turned out to be more descriptively accurate
than their knowledge-first alternative. While this is certainly a minor defeat for knowledge-first
epistemology, this explanatory failure does add up to a growing list of concepts that knowledge-
firsters struggle to handle within their framework
. As such, this paper has also contributed to make
a step (although a small one) towards reassessing the extent to which knowledge can be the game-
changing “unexplained explainer” that knowledge-firsters want it to be
Also Krauss (2017) proposes a definition that accounts for graded beliefs, but it is known to be vulnerable to
counterexamples (Benton 2018; Marsili 2019).
As I mentioned in footnotes 11 and 13, a minority of laypeople and philosophers have the intuition that an assertion
is a lie only if it is false, and this intuition can be accommodated by adding the further requirement that the proposition must
be false.
A growing body of literature, for instance, questions the possibility of analysing factive mental states (like seeing,
perceiving or remembering) in terms of knowledge (McDowell 2002; Bernecker 2009; Pritchard 2011; 2013; Turri 2010;
Schroeder manuscript; Cozzo 2010). For a broader overview of the obstacles encountered by the reductive ambitions of
knowledge-first epistemology, see McGlynn (2014) and Ger ken (2017).
The author would like to thank Vladimir Krstić, Michele Palmira, Giorgio Volpe and Alex Wiegmann for their helpful
feedback on earlier versions of this draft.
Forthcoming in Philosophical Studies
Augustine. DM. De Mendacio Liber Unus. Accessible at Wikisource, URL:
Aquinas. ST. Summa Theologiae. Accessible at Wikisource, URL:
Adler, Jonathan E. 1997. ‘Lying, Deceiving, or Falsely Implicating’. Journal of Philosophy 94 (9): 43552.
Arico, Adam J., and Don Fallis. 2013. ‘Lies, Damned Lies, and Statistics: An Empirical Investigation of the
Concept of Lying’. Philosophical Psychology 26 (6): 790816.
Benton, Matthew Aaron. 2017. ‘Lying , Belief , and Knowledge’. In The Oxford Handbook of Lying. 10.1093/oxfordhb/9780198736578.013.9.
Benton, Matthew Aaron. 2018. ‘Lying, Accuracy and Credence’. Analysis 78 (2): 19598.
Bernecker, Sven. 2009. Memory: A Philosophical Study. Oxford University Press.
Caminada, Martin. 2009. ‘Truth, Lies and Bullshit; Distinguishing Classes of Dishonesty’. In Proceedings of the
2009 IJCAI Workshop on Social Simulation.
Carson, Thomas L. 2006. ‘The Definition of Lying’. Noûs 2: 284306.
Carson, Thomas L. 2010. Lying and Deception. Oxford: Oxford University Press.
Chisholm, Roderick M., and Thomas D Feehan. 1977. ‘The Intent to Deceive’. Journal of Philosophy 74 (3):
Coignet, Martyn. 1586. Politique Discourses upon Trueth and Lying an Instruction to Princes to Keepe Their Faith and
Promise: Containing the Summe of Christian and Morall Philosophie, and the Duetie of a Good Man in Sundrie Politique
Discourses Vpon the Trueth and Lying.
Coleman, Linda, and Paul Kay. 1981. ‘Prototype Semantics: The English Word Lie’. Language 57 (1): 26.
Cozzo, Cesare. 2010. Is Knowledge the Most General Factive Stative Attitude? In Carlo Cellucci, Emiliano Ippoliti &
Emily Grosholz (eds.), Logic and Knowledge. Cambridge Scholars Press, 84-88.
Davis, Michael. 1986. ‘Why Attempts Deserve Less Punishment than Complete Crimes’. Law and Philosophy 5
(1): 1–32.
Dynel, Marta. 2018. Irony, Deception and Humour. Seeking the Truth about Overt and Covert Untruthfulness. Berlin,
Boston: De Gruyter, Mouton.
Fallis, Don. 2009. ‘What Is Lying?’ Journal of Philosophy 106 (1): 2956.
Fallis, Don. 2018. ‘What Is Deceptive Lying?’ In Lying: Language, Knowledge, Ethics, Politics. Oxford University
Gerken, Mikkel. 2017. ‘Against Knowledge-First Epistemology’. In Knowledge-First Approaches in Epistemology and
Mind, edited by Adam Carter, Emma Gordon, and Benjamin Jarvis, 46-71. Oxford: Oxford University Press.
Gupta, Anil. 2015. ‘Definitions’. In Stanford Encyclopedia of Philosophy (Summer 2015 edition).
Holguín, Ben. 2019. ‘Lying and Knowing’. Synthese, 198, 53515371.
Horn, Laurence R. 2017. ‘Telling It Slant: Toward a Taxonomy of Deception’. In The Pragmat ic Turn in Law:
Inference and Interpretation in Legal Discourse, 127.
Isenberg, Arnold. 1964. ‘Deontology and the Ethics of Lying’. Philosophy and Phenomenological Research 24 (4):
Kant, Immanuel. 1797. On a Supposed Right to Lie Because of Philantropic Concerns, translation of "Über ein
vermeintes Recht aus Menschenliebe zu lügen", Berlinische Blätter 10, 301314.
Krauss, Sam Fox. 2017. ‘Lying, Risk and Accuracy’. Analysis 73: 65159.
Krstić, Vladimir. 2019. ‘Can You Lie Without Intending to Deceive?’ Pacific Philosophical Quarterly 100 (2): 642
Kupfer, Joseph. 1982. ‘The Moral Presumption against Lying’. The Review of Metaphysics 36 (1): 10326.
Lackey, Jennifer. 2013. ‘Lies and Deception: An Unhappy Divorce’. Analysis 73 (2): 23648.
Leonard, HS. 1959. ‘Interrogatives, Imperatives, Truth, Falsity and Lies’. Philosophy of Science 26 (3): 17286.
Lindley, T Foster. 1971. ‘Lying and Falsity’. Australasian Journal of Philosophy, 49 (2): 3741.
Littlejohn, Clayton. 2009. ‘The New Evil Demon Problem’. In Internet Encyclopedia of Philosophy.
Mahon, James Edwin. 2015. ‘The Definition of Lying and Deception’. In Stanford Enciclopedia of Philosophy (Fall
2015 edition).
Mannison, D.S. 1969. ‘Lying and Lies’. Australasian Journal of Philosophy 2 (47): 13244.
Marsili, Neri. 2014. ‘Lying as a Scalar Phenomenon’. In Certainty-Uncertainty – and the Attitudinal Space in Between,
edited by Sibilla Cantarini, Werner Abraham, and Elisabeth Leiss, 153–73. Amsterdam: John Benjamins
Publishing Company.
Marsili, Neri. 2018a. ‘Lying and Certainty’. In The Oxford Handbook of Lying, edited by Jörg Meibauer, 169–82.
Oxford: Oxford University Press.
Marsili, Neri. 2018b. ‘Truth and Assertion: Rules versus Aims’. Analysis 78 (4): 63848.
Marsili, Neri. 2019. ‘Immoral Lies and Partial Beliefs’. Inquiry.
Marsili, Neri. 2020. ‘Lying, Speech Acts, and Commitment’. Synthese.
Marsili, Neri, and Alex Wiegmann. 2021. ‘Should I Say That? An Experimental Investigation of the Norm of
Assertion.’ Cognition, 212,
McDowell, John. 2002. ‘Responses’. In Reading McDowell: On Mind and World, edited by Nicholas H.Smith,
McGlynn, Aidan. 2014. Knowledge First? Palgrave Macmillian.
Meibauer, Jörg. 2014. Lying at the Semantics-Pragmatics Interface. Lying at the Semantics-Pragmatics Interface. Berlin,
Boston: De Gruyter.
Nelkin, Dana K. 2021. ‘Moral Luck’. In The Stanford Encyclopedia of Philosophy (Summer 2021 edition).
Pagin, Peter. 2016. ‘Problems with Norms of Assertion’. Philosophy and Phenomenological Research 93 (1): 178
Pritchard, Duncan. 2005. Epistemic Luck. Oxford: Clarendon Press.
Pritchard, Duncan. 2011. ‘Epistemological Disjunctivism and the Basis Problem’. Nous-Supplement: Philosophical
Issues 21 (1): 434–55.
Pritchard, Duncan. 2013. Epistemological Disjunctivism. Oxford: Oxford University Press.
Reins, Louisa M., and Alex Wiegmann. 2021. ‘Is Lying Bound to Commitment? Empirically Investigating
Deceptive Presuppositions, Implicatures, and Actions’. Cognitive Science 45.
Sartre, Jean-Paul. 1956. Being and Nothingness: An Essay on Phenomenological Ontology. Washington Press.
Saul, Jennifer. 2000. ‘Did Clinton Say Something False?’ Analysis 60 (3): 25557.
Saul, Jennifer. 2012. Lying, Misleading, and The Role of What Is Said. Oxford: Oxford University Press.
Schroeder, Mark. manuscript. ‘Knowledge Is Not the Most General Factive Stative Attitude’.
Siegler, Frederick A. 1966. ‘Lying’. American Philosophical Quarterly 3 (2): 12836.
Sosa, Ernest. 1999. ‘How to Defeat Opposition to Moore’. Philosophical Perspectives 33 (13): 14153.
Stokke, Andreas. 2018. Lying and Insincerity. Oxford: Oxford University Press.
Trpin, Borut, Anna Dobrosovestnova, and J G Sebastian. 2020. ‘A Computer Simulation Study of Graded Lies
and Trust Dynamics’. Synthese.
Turri, Angelo, and John Turri. 2015. ‘The Truth about Lying.’ Cognition 138: 16168.
Turri, Angelo, and John Turri. 2019. ‘Lying, Fast and Slow’. Synthese.
Turri, John. 2010. ‘Does Perceiving Entail Knowing?’ Theoria 76 (3): 197206.
Turri, John. 2016. Knowledge and the Norm of Assertion.
Turri, John. 2017. ‘Experimental Work on the Norms of Assertion’. Philosophy Compass 12 (7): 19.
Weissman, Benjamin, and Marina Terkourafi. 2018. ‘Are False Implicatures Lies? An Experimental
Investigation’. Mind & Language, 34, 221246.
Whyte, Jamie. 2013. ‘Review of “Lying, Misleading & What Is Said”, by J. M. Saul’. The Philosophical Quarterly,
Wiegmann, Alex, and Jörg Meibauer. 2019. ‘The Folk Concept of Lying’. Philosophy Compass, no. April: 111.
Wiegmann, Alex, Ronja Rutschmann, and Pascale Willemsen. 2017. ‘Empirically Investigating the Concept of
Lying’. Journal of Indian Council of Philosophical Research 34 (3): 591–609.
Wiegmann, Alex, Jana Samland, and Michael R. Waldmann. 2016. ‘Lying despite Telling the Truth’. Cognition
150: 37–42.
Wiegmann, Alex. forthcoming. True lies and Moorean redundancy. Synthese.
Williams, Bernard Arthur Owen. 2002. Truth and Truthfulness An Essay in Genealogy. Princeton: Princeton
University Press.
Williamson, Timothy. 2000. Knowledge and Its Limits. Oxford: Oxford University Press.
... 17 See Fallis (2010) for a defense of the methodological approach of conceptual analysis to the definition of lying. See also Marsili (2021) for a recent defense of conceptual analysis against rival methodologies in the definition of lying. 18 See Machery (2017) for arguments against conceptual analysis from experimental results that question the reliability of philosophical intuitions about cases. ...
... For example, Williamson, discussing the analysis of knowledge, claims that "The pursuit of analyses is a degenerating research programme" (2000, p. 31). 19 Compare my approach to Marsili (2021), who argues against KDL by presenting a case based on New Evil Demon scenarios. Marsili explicitly endorses the methodology of conceptual analysis and characterizes the methodology of theoretical adequacy as inconclusive. ...
Full-text available
The concept of lying is generally assumed to be closely related to the concept of assertion. However, the literature on lying has focused almost exclusively on lies expressed by unqualified assertions. Sometimes a speaker chooses to qualify her assertion by hedging, making her utterance a hedged declarative. This paper defends the thesis that lies can be expressed by untruthful hedged declaratives, and explores the implications of this thesis for the definition of lying. Many standard approaches to the definition of lying allow lies to be expressed by untruthful hedged declaratives, which may explain why these cases have been ignored in the literature. However, untruthful hedged declaratives present a problem for a recently-proposed theory that analyzes lying in terms of the Knowledge Norm of Assertion. According to this theory, a necessary condition on lying is that in asserting that p, the speaker represents herself as knowing that p; however, in making an untruthful hedged declarative a speaker represents herself as having an epistemic attitude weaker than knowledge. The fact that such a theory fails to adequately explain untruthful hedged declaratives provides a reason to prefer alternative theories of lying. The discussion draws out some important observations about the concepts of lying, assertion, belief, and the relations between them.
Suppose you assert a proposition p that you falsely believe to be false with the intention to deceive your audience. The standard view has it that you lied. This paper argues against orthodoxy: deceptive lying requires that p be in actual fact false, in addition to your intention to deceive by means of untruthfully asserting that p . We proceed as follows. First, an argument is developed for such falsity condition as the non‐psychological component of lying. The problem with the standard view, we profess, is exactly that lying is a purely psychological relation between disbelief, assertion, and intention. Then, by scrutinising familiar cases, we revisit the alleged intuitive support for the existence of true lies. It turns out these intuitions can be explained away once we reflect on the characteristic deceptive hallmarks that are associated with the distinction between lying and botched attempts at lying. Finally, we examine the morality of lying in the light of said falsity condition. The resultant view emphasises our moral sensitivity to the practical consequences of acts of lying, while still accommodating those moral considerations that pertain exclusively to the psychological components of lying.
Full-text available
Groups, like individuals, can communicate. They can issue statements, make promises, give advice. Sometimes, in doing so, they lie and deceive. The goal of this paper is to offer a precise characterisation of what it means for a group to make an assertion and to lie. I begin by showing that Lackey's influential account of group assertion is unable to distinguish assertions from other speech acts, explicit statements from implicatures, and lying from misleading. I propose an alternative view, according to which a group asserts a proposition only if it explicitly presents that proposition as true, thereby committing to its truth. This proposal is then put to work to define group lying. While scholars typically assume that group lying requires (i) a deceptive intent and (ii) a belief in the falsity of the asserted proposition, I offer a definition that drops condition (i) and significantly broadens condition (ii).
Full-text available
Stratified properties such as ‘happy unhappiness’, ‘ungrounded ground’, ‘fortunate misfortune’, and evidently ‘true falsity’ may generate dialetheias (true contradictions). The aim of the article is to show that if this is the case, then we will have a special, conjunctive , kind of dialetheia: a true state description of the form ‘Fa and not Fa’ (for some property F and object a), wherein the two conjuncts, separately taken, are to be held untrue. The particular focus of the article is on happy unhappiness: people suffering from (or enjoying) happy unhappiness (if there is some situation or state of mind of this kind) cannot be truly said ‘happy’ or ‘unhappy’, but we can say they are both. In the first section three cases of conjunctive stratification are presented; in the second section the logic of stratified contradictions is explored. The last section focuses on eudemonistic ascriptions: stated that a is happy to be unhappy (or unhappy to be happy), should we say a is happy? unhappy? both? neither?
Full-text available
Speech acts are governed by a variety of illocutionary norms. This chapter attempts to develop a common framework to study them, building on Sbisà’s (2019) work. Four families of illocutionary rules are identified: (i) Validity rules set conditions for (actual) performance; (ii) Cooperative rules set conditions for cooperative performance; (iii) Illocutionary goals set conditions for successful performance; (iv) Illocutionary obligations set conditions for compliance. Illocutionary rules are often taken to play a constitutive role: speech acts are said to be constituted by the unique set of rules that regulates them. Against this view, it is argued that many illocutionary rules are instead best construed as rationally derivable expectations of cooperation. This alternative paradigm provides fertile ground to reconcile disagreement between speech act theorists, and yields a promising explanation of how illocutionary norms are learned and evolve through time.
Full-text available
According to the subjective view of lying, speakers can lie by asserting a true proposition, as long as they believe this proposition to be false. This view contrasts with the objective view, according to which lying requires the actual falsity of the proposition asserted. The aim of this paper is to draw attention to pairs of assertions that differ only in intuitively redundant content and to show that such pairs of assertions are a reason to favour the subjective view of lying over the objective one.
Full-text available
Assertions are our standard communicative devices for sharing and acquiring information. Recent studies seemingly provide converging evidence that assertions are subject to a factive norm: you are entitled to make an assertion only if it is true. However, these studies assume that we can treat participants' judgements about what an agent ‘should say’ as evidence of their intuitions about assertability. This paper argues that this assumption is incorrect, so the conclusions drawn in the aforementioned studies are unwarranted. We provide evidence that most people do not interpret statements about what one ‘should say’ as statements about assertability, but rather as statements about what is in the agent's interest to do. Measures for prompting the intended reading of the test question are identified, and their efficacy is tested. We found that when these measures are implemented, people's judgements consistently and overwhelmingly align with non-factive accounts of assertion.
Full-text available
Lying is an important moral phenomenon that most people are affected by on a daily basis—be it in personal relationships, in political debates, or in the form of fake news. Nevertheless, surprisingly little is known about what actually constitutes a lie. According to the traditional definition of lying, a person lies if they explicitly express something they believe to be false. Consequently, it is often assumed that people cannot lie by more indirectly communicating believed‐false claims, for instance by merely conversationally implicating them. In this paper, we subject this claim to an empirical test. In a preregistered study of 300 participants, we investigate how people judge cases of implicit deceptions that would usually be excluded by the traditional definition of lying (i.e., conversational implicatures, presuppositions, and nonverbal actions). Our results show that people do in fact consider it possible to lie by indirect means, suggesting that people have a broader concept of lying than is usually assumed. Moreover, our findings indicate that lie judgments are closely tied to the extent to which agents are perceived as having committed themselves to the believed‐false claims they have communicated. We discuss the implications of our results for the traditional definition of lying and propose a new commitment‐based definition of lying that can account for the findings of our experiment.
Full-text available
Not every speech act can be a lie. A good definition of lying should be able to draw the right distinctions between speech acts (like promises, assertions, and oaths) that can be lies and speech acts (like commands, suggestions, or assumptions) that under no circumstances are lies. This paper shows that no extant account of lying is able to draw the required distinctions. It argues that a definition of lying based on the notion of ‘assertoric commitment’ can succeed where other accounts have failed. Assertoric commitment is analysed in terms of two normative components: ‘accountability’ and ‘discursive responsibility’. The resulting definition of lying draws all the desired distinctions, providing an intensionally adequate analysis of the concept of lying.
Full-text available
Partial lying denotes the cases where we partially believe something to be false but nevertheless assert it with the intent to deceive the addressee. We investigate how the severity of partial lying may be determined and how partial lies can be classified. We also study how much epistemic damage an agent suffers depending on the level of trust that she invests in the liar and the severity of the lies she is told. Our analysis is based on the results from exploratory computer simulations of an arguably rational Bayesian agent who is trying to determine how biased a coin is while observing the coin tosses and listening to a (partial) liar's misleading predictions about the outcomes. Our results provide an interesting testable hypothesis at the intersection of epistemology and ethics, namely that in the longer term partial lies lead to more epistemic damage than outright lies.
Researchers have debated whether there is a relationship between a statement’s truth-value and whether it counts as a lie. One view is that a statement being objectively false is essential to whether it counts as a lie; the opposing view is that a statement’s objective truth-value is inessential to whether it counts as a lie. We report five behavioral experiments that use a novel range of behavioral measures to address this issue. In each case, we found evidence of a relationship. A statement’s truth-value affects how quickly people judge whether it is a lie. When people consider the matter carefully and are told that they might need to justify their answer, they are more likely to categorize a statement as a lie when it is false than when it is true. When given options that inhibit perspective-taking, people tend to not categorize deceptively motivated statements as lies when they are true, even though they still categorize them as lies when they are false. Categorizing a speaker as “lying” leads people to strongly infer that the speaker’s statement is false. People are more likely to spontaneously categorize a statement as a lie when it is false than when it is true. We discuss four different interpretations of relevant findings to date. At present, the best supported interpretation might be that the ordinary lying concept is a prototype concept, with falsity being a centrally important element of the prototypical lie.