British Journal for the History of Philosophy 14(4) 2006: 653 – 685
ARTICLE
KANT AND THE PERFECT DUTY TO OTHERS
NOT TO LIE
James Edwin Mahon
INTRODUCTION
A great contempt for lies and liars is evident in all of Kant’s writings on
practical philosophy and theology.1 Despite the frequency and the virulence
1
References to Kant’s works in the text and endnotes are given parenthetically, according
to the abbreviations listed below. Pagination is as follows: first to the volume and page number
in the standard edition of Kant’s works, Kants gesammelte Schriften, edited under the auspices
of the Königliche Preussische Akademie der Wissenschaften (Berlin: Walter de Gruyter, 1908–
13) second to the page number in the respective translation. Unless otherwise indicated,
all emphases are in the original.
A: Anthropology from a Pragmatic Point of View (Anthropologie in pragmatischer Hinsicht)
(1798), translated by Mary J. Gregor (The Hague: Martinus Nijhoff, 1974).
AN: Announcement of the Near Conclusion of a Treaty for Eternal Peace in Philosophy
(Verkündigung des nahen Abschlusses eines Traktes zum ewigen Frieden in der Philosophie)
(1796), translated by Peter Fenves, in Raising the Tone of Philosophy: Late Essays by Immanuel
Kant, Transformative Critique by Jacques Derrida edited by Peter Fenves (Baltimore: The Johns
Hopkins University Press, 1993) 83–100.
C: Correspondence (Kant’s Briefwechsel), translated and edited by Arnulf Zweig (Cambridge:
Cambridge University Press, 1999).
CPR: Critique of practical reason (Kritik der praktischen Vernunft) (1788), translated by Mary J.
Gregor, in Practical Philosophy, translated and edited by Mary J. Gregor and Allen W. Wood
(Cambridge: Cambridge University Press, 1996) 133–271.
E: The Educational Theory of Immanuel Kant (Über Pädagogik) (1803) (edited by Friedrich
Theodor Rink), translated by Annette Churton (Ann Arbor, Michigan: University of Michigan
Press, 1960).
G: Groundwork of the Metaphysics of Morals (Grundlegung zur Metaphysik der Sitten) (1785),
translated by Mary J. Gregor, in Practical Philosophy, 37–108.
LE: Lectures on ethics (Vorlesungen über Ethik) (1924) translated by Peter Heath and edited by
Peter Heath and J. B. Schneewind (Cambridge: Cambridge University Press, 1997).
MM: The metaphysics of morals (Die Metaphysik der Sitten) (1797), translated by Mary J.
Gregor, in Practical Philosophy, 353–603.
M: On the miscarriage of all philosophical trials in theodicy (Über das Misslingen aller
philosophischen Versuche in der Theodizee) (1791), translated by George di Giovanni, in Religion
and Rational Theology, edited and translated by Allen W. Wood and George di Giovanni
(Cambridge: Cambridge University Press, 1996) 24–37.
British Journal for the History of Philosophy
ISSN 0960-8788 print/ISSN 1469-3526 online ª 2006 BSHP
http://www.tandf.co.uk/journals DOI: 10.1080/09608780600956407
654
JAMES EDWIN MAHON
with which lies and liars are condemned, however, and the fervour with which
truthfulness is advocated,2 Kant’s practical philosophy (the ethics and the
philosophy of right) leaves us with at least two important questions about the
duty to be truthful. Or, to be clearer, and to avoid any confusion with the duty
to be candid in his ethics,3 it leaves us with at least two important questions
about the duty not to lie. The first is whether the duty not to lie is an ethical
duty or a duty of right, or whether there is both an ethical duty not to lie and a
duty of right not to lie. The second is whether, if there is an ethical duty not to
lie, it is a duty to others or a duty to oneself, or whether there is both an ethical
duty to others not to lie and an ethical duty to oneself not to lie.
O: Observations on the Feeling of the Beautiful and Sublime (Beobachtungen über das Gefühl des
Schönen und Erhabenen) (1764), translated and edited by John T. Goldthwaite (Berkeley:
University of California Press, 1973).R: Religion within the boundaries of mere reason (Die
Religion innerhalb der Grenzen der blossen Vernunft) (1793), translated by George di Giovanni, in
Religion and Rational Theology, 39–215.
RL: On a supposed right to lie from philanthropy (Über ein vermeintes Recht aus Menschenliebe zu
lügen) (1797), translated by Mary J. Gregor, in Practical Philosophy, 611–15.
2
Lack of space prevents an exhaustive inventory of Kant’s invectives against liars and lying. In
Observations on the Feeling of the Beautiful and Sublime in 1764 Kant concurs with The
Spectator that ‘no more insulting reproach could be made to a man than if he is considered a
liar’ (O, 2 (233): 83). In the Critique of practical reason in 1788 Kant says that a person should
choose to die rather than ‘give false testimony against an honorable man’ (CPR, 5 (30): 163). In
On the miscarriage of all philosophical trials in theodicy in 1791 Kant says that the ‘principal
affliction of human nature’ is the propensity to ‘mendacity (falsity, even without any intention to
harm)’, and that a liar has the evil of ‘baseness, whereby all character is denied to the human
being’ (M, 8 (270): 29). In Religion within the boundaries of mere reason in 1793 Kant claims that
without ‘sincerity (that everything said be said with truthfulness)’, the ‘human race would have
to be in its own eyes an object of deepest contempt’ (R, 6 (190 n.2): 206 n.2). In his lectures on
ethics given in 1793–4 Kant says that ‘A man who lies demeans himself in the eyes of others, in
his testimony and judgement he loses all credibility; for society, he is to be regarded in all his
tales as a mute, since we never know how far he can be trusted’ (LE, 27 (605): 351). In the
Doctrine of Virtue, Part II of The Metaphysics of Morals, in 1797 (among other places) Kant
points out that ‘It is noteworthy that the Bible dates the first crime, through which evil entered
the world, not from fratricide (Cain’s) but from the first lie (for even nature rises up against
fratricide), and calls the author of all evil a liar from the beginning and the father of lies.’ (MM,
6 (431): 554) And in the lectures on pedagogy, published posthumously in 1803, Kant says that
‘truthfulness . . . is the foundation and very essence of character. A man who tells lies has no
character’ (E, 9 (484): 90). It is interesting to note that in all of these invectives Kant does not
distinguish between lies, or different circumstances in which lies are told. All lies are uniformly
and equally condemned. Arnulf Zweig has pointed out that ‘Kant spoke of his father as honest,
hard working, and deeply intolerant of lies – an intolerance that his son evidently internalized’
(‘A Sketch of Kant’s Life’, in Immanuel Kant: Groundwork for the Metaphysics of Morals,
translated by Arnulf Zweig and edited by Thomas E. Hill, Jr., and Arnulf Zweig (Oxford:
Oxford University Press, 2002) 8).
3
In addition to (as I argue) the two duties not to lie in Kant’s ethics, there is also the duty to
others to be candid to others (cf. MM, 6 (433 n.1): 556 n.1). Unlike the two duties not to lie,
which are perfect negative duties, the duty to be candid is an imperfect positive duty. For more
on the imperfect positive duty to others to be candid to others in Kant’s ethics, see my ‘Kant on
Lies, Candour and Reticence’, Kantian Review 7 (2003) 101–33.
KANT AND THE DUTY NOT TO LIE
655
In this article I shall concentrate on Kant’s ethics, and leave to one side
his philosophy of right.4 I believe that one can find two duties not to lie in
Kant’s ethics or, as I shall put it, two ethical duties not to lie. Both of these
ethical duties are exceptionless negative duties. The first is the perfect duty to
others not to lie to others. This ethical duty can be found in the lectures on
ethics in 1762–4, in the lectures on ethics in 1784–5, in the Groundwork of the
Metaphysics of Morals in 1785, and in the lectures on ethics in 1793–4. This
ethical duty is not referred to, however, in any of Kant’s publications after
1785. The second is the perfect duty to oneself not to lie to oneself or to others.
This ethical duty can be found in the lectures on ethics in 1762–4, in the
lectures on ethics in 1784–5, in the letter to Maria von Herbert in 1792, in
the lectures on ethics in 1793–4, and in the Doctrine of Virtue, part II of The
Metaphysics of Morals, in 1797. It is also briefly mentioned in a footnote in
On a supposed right to lie from philanthropy in 1797. This ethical duty is not
referred to in the Groundwork, however.
In this article I am concerned with the first exceptionless negative duty,
the perfect duty to others not to lie to others.5 I shall argue that one
argument for this duty, drawn from the Universal Law Formula of the
Categorical Imperative, fails to generate the requisite exceptionless
negative duty. This argument cannot prohibit a certain class of lies to
others, namely, lies told to liars (or at least, to those who are believed to
be liars). I shall also argue that another argument for this duty, drawn
from the Humanity as an End in Itself Formula of the Categorical
Imperative, also fails to generate the requisite exceptionless negative duty.
This argument also cannot prohibit a certain class of lies to others,
namely, lies told to those who have consented to being lied to, without
knowing which lies they are to be told. I shall conclude that no argument
drawn from the Categorical Imperative can generate a perfect duty to
others not to lie to others.
Before proceeding, an explanatory remark about the naming of these two
ethical duties is in order. Kant is normally concerned with lies that are told
to other persons. He does hold, however, that it is possible for a person to lie
to herself. The perfect duty to others not to lie does not prohibit lying to
oneself. It only prohibits lying to others.6 But the perfect duty to oneself not
to lie does prohibit lying to others.7 It prohibits lying to oneself and lying to
4
I will discuss lying in Kant’s philosophy of right on another occasion. I should say here
however that I believe that there are two duties of right not to lie in Kant’s philosophy of right.
There is the duty of right to others not to lie to others, and the duty of right to humanity in
general not to lie to others.
5
I will discuss the perfect duty to oneself not to lie to oneself or to others on another occasion.
6
Kant provides no argument as to why the perfect duty to others not to lie does not prohibit
lying to oneself.
7
Nelson Potter has commented on this:
In the Tugendlehre [Doctrine of Right] lying is classified as a violation of ‘The Human
Being’s Duty to Himself Merely as a Moral Being’. In fact, it is the greatest such
656
JAMES EDWIN MAHON
others. Thus the scope of the duty to others is narrower than the scope of the
duty to oneself. As a result, the first duty is properly named the perfect duty
to others not to lie to others. It is a duty to others, regarding others. The
second duty is properly named the perfect duty to oneself not to lie to
oneself or to others. It is a duty to oneself, regarding oneself and others. For
convenience, however, I shall sometimes refer to these two duties as the duty
to others not to lie and the duty to oneself not to lie.
THE PERFECT DUTY TO OTHERS NOT TO LIE
TO OTHERS PRIOR TO THE GROUNDWORK: THE EARLY
LECTURES ON ETHICS
In his lectures on ethics in 1762–4 (as recorded by Johann Gottlieb Herder)
Kant says that the definition of a lie in the philosophy of right, according to
which a lie is an untruthful declaration that harms another person (what he
elsewhere in his lectures calls a ‘falsiloquium in praejudicium humanitatis’
[falsehood that damages another person] (LE, 27: 448 (p. 203)), is too
narrow for ethics. In ethics an untruthful declaration is a lie, and is ethically
impermissible, regardless of whether or not another person is harmed:
‘Lying is simply too restricted, as an injury to the other; as untruth, it
already has an immediately abhorrent quality’ (LE, 27: 59 (p. 25)). It is
important to point out that Kant makes the same claim about lies many
years later in the Doctrine of Virtue:
In the doctrine of right an intentional untruth is called a lie only if it violates
another’s right; but in ethics, where no authorization is derived from
harmlessness, it is clear of itself that no intentional untruth in the expression
of one’s thoughts can refuse this harsh name.
(MM, 6: 429 (p. 552))
In these early lectures on ethics Kant argues that being truthful, or not lying,
is both an ethical duty to others (a duty ‘of virtue’ to others) and an ethical
duty to oneself (a duty ‘of virtue’ to oneself):
Value of the love of truth: It is the basis of all virtue; the first law of nature, Be
truthful!, is a ground (1) of virtue towards others, for if all are truthful, a man’s
untruth would be exposed as a disgrace. (2) of virtue to oneself, for a man
cannot hide from himself, not is he able to contain his abhorrence.
(LE, 27: 60 (p. 26))
violation . . . This classification has had a certain amount of attention from scholars,
since more commonly, and perhaps more intuitively, lying is considered a violation of
duty to the person lied to.
(‘Duties to Oneself in Kant’s Ethics’, in Kant’s Metaphysics of Morals: Interpretive
Essays, edited by Mark Timmons (Oxford: Oxford University Press, 2002) 386)
KANT AND THE DUTY NOT TO LIE
657
Here Kant criticizes Alexander Gottlieb Baumgarten, whose textbooks on
practical philosophy Kant used in his lectures,8 for arguing that a lie only
violates an ethical duty to oneself. A lie, or at least a lie told to another
person, also violates an ethical duty to others: ‘Certain untruths are not
called lies, because the latter are strictly untruths that are contrary to duty;
not, however, as our author [Baumgarten] thinks, merely to the duty to
myself, but also to that towards others’ (LE, 27: 61 (p. 27)).
In his lectures on ethics in 1784–5 (as recorded by Georg Ludwig
Collins), under the heading ‘Of Duties to Oneself ’, Kant says that ‘The
lie is more an infringement of duty to oneself than to others, and even if
a liar does nobody any harm by it, he is still an object of contempt, a
low fellow who violates the duties to himself.’ (LE, 27: 341 (p. 123)) This
statement is ambiguous. It can be read as stating that a lie to another
person that does not harm the other person only violates an ethical duty
to oneself. If so, then Kant is not here defending an ethical duty to
others not to lie. Alternatively, it can be read as stating that a lie to
another person that does not harm the other person does indeed violate
an ethical duty to others, as well as an ethical duty to oneself. If this is
so, then Kant is here defending an ethical duty to others not to lie, and
Kant should clearly state ‘and even if a liar does nobody any harm by it,
he is still a low fellow who violates the duty to others, and the duty to
himself ’.
Later in these same lectures, under the general heading ‘Of Duties
Towards Other People’ (LE, 27: 413 (p. 177)), Kant discusses lying and
truthfulness under the heading ‘Of Ethical Duties Towards Others, and
Especially Truthfulness’ (LE, 27: 444 (p. 200)). Here he says that ‘In human
social life, the principal object is to communicate our attitudes, and hence it
is of the first importance that everyone be truthful in respect of his thoughts,
since without that, social intercourse ceases to be of any value.’ (LE, 27: 444
(p. 200–1)) He does not, in his further discussion of being truthful, explain
why others are owed the duty to be truthful. He does however insist that ‘If
everyone were well disposed, it would not only not be a duty not to lie, but
nobody would need to do it, since he would have nothing to worry about.’
(LE, 27: 448 (p. 204)) The first part of this claim, however, seems false. Even
if everyone were well disposed to everyone else, it would still be a duty to
others not to lie to others.
8
Kant used Baumgarten’s Introduction to Practical First Philosophy (Initia philosophiae
practicae primae) (3rd edn, 1760) and his Philosophical Ethics (Ethica philosophica) (1st edn,
1751, and 2nd edn, 1763) as textbooks in his lectures on ethics (he was required by law to use
textbooks). Baumgarten’s two Latin textbooks are reprinted in Vols 19 and 27, respectively, of
the Königliche Preussische Akademie der Wissenschaften edition of the works of Immanuel
Kant.
658
JAMES EDWIN MAHON
THE PERFECT DUTY TO OTHERS NOT TO LIE TO OTHERS IN
THE GROUNDWORK: THE PREFACE AND SECTION I
The very first moral law stated in the Groundwork is ‘thou shalt not lie’ (‘du
sollst nicht lügen’). This law is given in the Preface. As with all moral laws, it
applies to all rational beings, and not merely to all human beings. Or at
least, it applies to all rational beings who are capable of lying.9
Everyone must grant that a law, if it is to hold morally, that is, as a ground of
an obligation, must carry with it absolute necessity; that, for example, the
command ‘thou shalt not lie’ does not hold only for human beings, as if other
rational beings did not have to heed it, and so with all other moral laws
properly so called.
(G, 4: 389 (p. 44–5))
Following the Preface, in Section I of the Groundwork Kant gives an early
statement of what he later (in Section II) calls the Categorical Imperative, in
its Universal Law Formula: ‘I ought never to act except in such a way that I
could also will that my maxim should become a universal law’ (G, 4: 402
(p. 57)). Here he immediately moves to consider the question ‘may I, when
hard pressed, make a promise with the intention not to keep it?’ (G, 4: 402
(p. 57)) Thus the very first example of a maxim of action considered in the
Groundwork is the maxim of lying to gain money.
Kant argues here that such a maxim cannot be made into a universal
law.10 That is, as one commentator has put it, the maxim cannot ‘universally
determine everyone’s conduct, and hence . . . become everyone’s principle of
action’.11 There cannot be a universal law of gaining money by means of
making a lying promise to repay money if it is given. This is because, if there
were a universal law according to which everyone who wanted money made
a lying promise to repay money if it were given to her, then those in the
position to lend money, knowing this law, and not being naturally disposed
to give away their money, would not lend money to anyone who promised
to repay money if it were given to her. They would refuse, or lie in return,
because of the possibility that the person who promised to repay the money
was following this law and was lying, and because they are not naturally
9
In the Anthropology Kant considers the following possible rational beings: ‘It could well be
that some other planet is inhabited by rational beings who have to think aloud—who, whether
awake or dreaming, in company with others or alone, can have no thoughts they do not utter.’
(A, 7 (333) 192) Such rational beings would not be capable of lying. Hence the moral law ‘Thou
shalt not lie’ would fail to apply to them.
10
As will become apparent below, I favour the interpretation of the Formula of Universal Law
according to which a universal law is a law of nature. Further, I favour the interpretation of a
law of nature according to which a law of nature is a causal law.
11
Leslie Mulholland, ‘Kant: On willing maxims to become laws of nature’, Dialogue 18
(1978) 95.
KANT AND THE DUTY NOT TO LIE
659
12
disposed to give handouts to others. Hence, no one who wanted money
would ever gain money by means of making a lying promise. Making a lying
promise to gain money would result in the person not gaining money.
Making a lying promise to gain money would be futile.13 Since the ‘form of
lawfulness is the form of order or regularity in the production of an effect’,14
and since the effect (gaining money) would not be produced by the cause
(making a lying promise), it follows that there cannot be a universal law of
gaining money by this means:
However, to inform myself in the shortest and yet infallible way about the
answer to this problem, whether a lying promise is in conformity with duty, I
ask myself: would I indeed be content that my maxim (to get myself out of
difficulties by a false promise) should hold as a universal law (for myself as well
as for others)? and could I indeed say to myself that every one may make a
false promise when he finds himself in a difficulty he can get out of in no other
way? Then I soon become aware that I could indeed will the lie, but by no
means a universal law to lie; for in accordance with such a law there would
properly be no promises at all, since it would be futile to avow my will with
regard to my future actions to others who would not believe this avowal or, if
they rashly did so, would pay me back in like coin; and thus my maxim, as
soon as it were made a universal law, would have to destroy itself.
(G, 4: 403 (p. 57))
12
Kant does assume that, among other things, it is a contingent truth about people that they are
not naturally disposed to give handouts to others. As Shelly Kagan has pointed out: ‘Regardless
of which interpretation we accept, it is worth drawing attention to the fact that the argument
makes use of various contingent, empirical facts. The argument assumes, for example, that
people have memories’ (‘Kantianism for Consequentialists’, in Immanuel Kant, Groundwork for
the Metaphysics of Morals, edited and translated by Allen W. Wood (New Haven: Yale
University Press, 2002) 131).
13
Mulholland argues that Kant’s argument is that making promises would be impossible, and
hence, that making a lying promise to obtain money would be impossible, and hence, that
obtaining money by means of making a lying promise would be impossible:
given that no one believed what was promised him, it follows that it would be
impossible to make promises. Promises are made only if they are accepted. But
promises are accepted only if they are believed. Hence, if no one believed what was
promised him, promises could not be made.
(Mulholland, 1978: 98)
My argument is the weaker one that Kant’s argument is that obtaining money, by means of
making a lying promise, would be impossible, since lying promises would not be believed. I do
not argue that Kant argues that making a lying promise to obtain money would be impossible,
much less that he argues that making promises would be impossible. Here I agree with Allen
Wood that
it is far from self-evident that if M2 [the maxim] were a universal law of nature then
there could be no promises at all. It is much easier to convince oneself that U2 [the
universalization of the maxim] as a universal law of nature would simply render it
impossible to achieve one’s ends by making false promises.
(Kant’s Ethical Thought (Cambridge: Cambridge University Press, 1999: 89)
14
Mulholland, 1978: 95.
660
JAMES EDWIN MAHON
One commentator has argued that ‘What Kant in fact supposes to be
universalized is not a law but a permission – ‘Everyone may make a deceitful
promise when he finds himself in a difficulty from which he cannot otherwise
extricate himself.’ This would, no doubt, lead to chaos. Then we should
never know whether the maker of the promise meant to keep it or to break
it’.15 The reason for saying that Kant attempts to universalize a permission,
and not a law, is that ‘The maxim, ‘People, when they make promises, shall
always make promises they do not mean to fulfil’, is not one which would
necessarily destroy itself. For if people acted always on this maxim every
promise could be relied upon to be broken, and this would in many (though
not in all) cases serve as well as if it could be relied on to be kept.’16
However, this is wrong. Kant does indeed suppose a law to be universalized,
and not merely a permission. The chaos that would ensue if everyone who
needed money were to follow this law would stem from the uncertainty of
those in the position to lend money as to whether the person was one of
those following this law, or whether the person was someone who only
needed a loan (as it were, someone following a different law, namely, make a
truthful promise to repay money if it is given in order to borrow money). No
amount of questioning could settle this matter, of course. This uncertainty,
coupled with the fact that they are not naturally disposed to give handouts
to others, would lead to their refusal or lying in return.
Since the maxim of gaining money by means of making a lying promise to
repay money if it is given cannot be made into a universal law, it follows that
acting on the maxim, that is, making a lying promise to repay money if it is
given in order to gain money, is ethically impermissible. Kant’s conclusion
in Section I of the Groundwork, however, is not merely that making a lying
promise in order to gain money is ethically impermissible. His conclusion is
that lying for any purpose whatsoever is ethically impermissible. That is, his
conclusion is that lying, as such, is ethically impermissible, and that one has
a ethical duty ‘To be truthful’:
To be truthful from duty, however, is something entirely different from being
truthful from anxiety about detrimental results, since in the first case the concept
of the action in itself already contains a law for me while in the second I must first
look about elsewhere to see what effects on me might be combined with it.
(G, 4: 402 (p. 57))
Although Kant does say in these early parts of the Groundwork that ‘thou
shalt not lie’ is a moral law for all rational beings capable of lying, and
15
W. D. Ross, Kant’s Ethical Theory: A Commentary on the Grundlegung zur Metaphysik der
Sitten (Oxford: Clarendon Press, 1954) 30. A similar interpretation has been offered by Thomas
W. Pogge in ‘The Categorical Imperative’, Grundlegung der Metaphysik der Sitten: Ein
kooperativer Kommentar, edited by Otfried Höffe (Frankfurt am Main: Vittorio Klostermann,
1989) 172–93.
16
Ibid.: 30.
KANT AND THE DUTY NOT TO LIE
661
that there is an ethical duty ‘To be truthful’, he does not explicitly
characterize this ethical duty as a duty to others. Presumably this is because
he is not, in these early parts of the Groundwork, engaged in characterizing
ethical duties.
THE PERFECT DUTY TO OTHERS NOT TO LIE
TO OTHERS IN THE GROUNDWORK: SECTION II
In Section II of the Groundwork, Kant says that he will enumerate ethical
‘duties to ourselves and to other human beings’, which he further divides
into ‘perfect and imperfect duties’ to ourselves, and perfect and imperfect
duties to others (G, 4: 421 (p. 73)). However, in enumerating these ethical
duties, by way of examples, Kant does not state in so many words which
particular ethical duties he has in mind. In particular, he does not state in so
many words which particular perfect duty to others he has in mind.
Nevertheless, I shall argue that the perfect duty to others he has in mind is
the perfect duty to others not to lie to others.
In providing an example of a perfect duty to others in Section II of the
Groundwork, Kant returns to the case described in Section I, that of a person
‘who has it in mind to make a false promise to others’ (G, 4: 429 (p. 80)). The
‘maxim of action’ (G, 4: 422 (p. 74))17 that Kant considers is, ‘when I believe
myself to be in need of money I shall borrow money and promise to repay it,
even though I know that this will never happen.’ (G, 4: 422 (p. 74)). Since, if
one does indeed borrow money, it follows that one does, implicitly or
explicitly, promise to repay it, this maxim is awkwardly expressed. It would
be better to say ‘when I believe myself to be in need of money I shall take
money and promise to repay it, even though I know that this will never
happen’, or ‘when I believe myself to be in need of money I shall take money
and lie that I am borrowing it’.
The argument here is the same as that given in Section I of the
Groundwork. It is drawn from the ‘categorical imperative’, in its Universal
Law Formula: ‘act only in accordance with that maxim through which you can
at the same time will that it become a universal law’ (G, 4: 421 (p. 73)). The
maxim ‘when I believe myself to be in need of money I shall take money and
promise to repay it, even though I know that this will never happen’, cannot
be made into a universal law. If everyone who needed money were to make a
17
In the Groundwork Kant considers two kinds of maxim: maxims of action and maxims of
ends. Maxims of action, such as ‘in order to gain money, I shall lie that I am borrowing money’,
involve actions (here, the action of lying). Maxims of ends, such as ‘in order to enjoy myself,
I shall neglect my natural gifts’, involve ends (here, neglecting one’s natural gifts). Many
different actions may satisfy the end of neglecting one’s natural gifts. Consideration of maxims
of action typically leads to perfect duties (e.g. the duty to others not to lie). Consideration of
maxims of ends typically leads to imperfect duties (e.g. the duty to oneself to develop one’s
natural gifts).
662
JAMES EDWIN MAHON
lying promise to repay money if it were given to her, then no one would ever
gain money by this means. No one in the position to lend money, knowing
that people who need money lie that they will repay it, and not being
naturally disposed to give handouts to others, would ever lend money to
anyone who asked to borrow it. They would ‘laugh’ at people who asked to
borrow money:
For, the universality of a law that everyone, when he believes himself to be in
need, could promise whatever he pleases with the intention of not keeping it
would make the promise and the end one might have in it itself impossible,
since no one would believe what was promised him but would laugh at all such
expressions as vain pretences.
(G, 4: 422 (p. 74))
Since the maxim of gaining money by means of making a lying promise to
repay money if it is given cannot be made into a universal law, it follows that
acting on the maxim, that is, making a lying promise to repay money if it is
given in order to gain money, is ethically impermissible. As he also says in
Section II:
For example, when it is said ‘you ought not to promise anything deceitfully,’
and one assumes that the necessity of this omission is not giving counsel for
avoiding some ill or other – in which case what is said would be ‘you ought not
to make a lying promise lest if it comes to light you destroy your credit’ – but
that an action of this kind must be regarded as in itself evil and that the
imperative of prohibition is therefore categorical.
(G, 4: 419 (pp. 71–2))
Unlike in Section I, however, in Section II Kant does not say that the perfect
duty to others in question in this example is the perfect duty to others to be
truthful, that is, the perfect duty to others not to lie. Nevertheless, I believe
that the perfect duty to others at issue here is the perfect duty to others not
to lie to others.
I shall return to the argument used to generate this perfect duty to others.
Before doing so, however, I want to examine passages in the Critique of
Practical Reason, and in the later lectures on ethics, that have a bearing on
this argument.
THE PERFECT DUTY TO OTHERS NOT TO LIE
TO OTHERS AFTER THE GROUNDWORK: THE CRITIQUE
OF PRACTICAL REASON
The Critique of Practical Reason contains little discussion of duties in
general, and little discussion of the duty not to lie in particular. Kant does
KANT AND THE DUTY NOT TO LIE
663
provide, however, as an example of a practical law or categorical imperative,
the rule ‘never to make a lying promise’:
Now tell someone that he ought never to make a lying promise; this is a rule
that has to do only with his will, regardless of whether the purposes the human
being may have can be thereby attained; the mere volition is that which is to be
determined completely a priori by this rule. If, now, it is found that this rule is
practically correct, then it is a law because it is a categorical imperative.
(CPR, 5: 21 (p. 154–5))
One possible implication of this passage of the Critique is that the duty to
others in Section II of the Groundwork is the perfect duty to others not to
make lying promises to others, as opposed to the perfect duty to others not
to lie to others. However, later in the Critique Kant summarizes the
arguments of the second section of the Groundwork. Here he argues that a
maxim of deceiving when it is believed to be to one’s advantage cannot be
universalized. The implication is that the perfect duty to others that is at
issue in Section II of the Groundwork is not merely the perfect duty to others
not to make lying promises to others. It is, rather, the perfect duty to others
not to deceive others.
Thus one says: if everyone permitted himself to deceive when he believed it to
be to his advantage, or considered himself authorized to shorten his life as
soon as he was thoroughly weary of it, or looked with complete indifference on
the need of others, and if you belonged to such an order of things, would you
be in it with the assent of your will? Now everyone knows very well that if he
permits himself to deceive secretly it does not follow that everyone else does so,
or that if, unobserved, he is hard-hearted everyone would not straightaway be
so towards him; accordingly, this comparison of the maxim of his actions with
a universal law of nature is also not the determining ground of his will.
(CPR, 5: 69 (p. 196))
Even here, however, Kant does not come out and say that there is a perfect
duty to others not to deceive others. Later in the Critique, when discussing
the teaching of virtue to children, Kant chooses the case of someone who is
offered bribes, threatened by powerful figures, and implored by friends and
family, to lie that Anne Boleyn is guilty (of adultery or sedition or some such
crime), but who ‘remains firm in his resolution to be truthful, without
wavering or even doubting’ (CPR, 5: 156 (p. 264–5)). He also quotes a
favourite passage from Juvenal’s eighth Satire: ‘if summoned to bear witness
in some dubious and uncertain cause, though Phalaris himself should dictate
that you perjure yourself and bring his bull to move you, count it the
greatest of all iniquities to prefer life to honor and to lose, for the sake of
living, all that makes life worth living.’ (CPR, 5: 158–9, p. 267 n.1). Despite
what is said here and in the earlier parts of the Critique about deception and
664
JAMES EDWIN MAHON
truthfulness, however, there is no statement in this work to the effect that
there is a perfect duty to others not to lie to others.
THE PERFECT DUTY TO OTHERS NOT
TO LIE TO OTHERS AFTER THE GROUNDWORK:
THE LATER LECTURES ON ETHICS
In his lectures on ethics in 1793–4 (as recorded by Johann Friedrich
Vigilantius) Kant argues that there is an ‘imperative of reason’ to ‘absolutely
speak the truth’, or to not lie. He does so by way of an argument similar to
that given in the Groundwork, namely, that a maxim of lying in order to gain
a great advantage cannot be made into a universal law. Importantly, Kant
claims that this argument establishes that there is a ‘perfect’ duty to
‘absolutely speak the truth’:
You shall absolutely speak the truth, is an imperative of reason, and in
application a maxim which reason converts into a universal law. For suppose
that someone were to have the maxim, that he might tell an untruth whenever
he could thereby obtain great advantage, the question arises, whether this
maxim could stand as a universal law. We would then have to presuppose that
nobody will tell the truth to his disadvantage, and in that case nobody would
continue to have any trust; the liar could thus never succeed in deceiving
anyone by lying, and the law would therefore automatically destroy itself. So it
is with all perfect duties; if the opposite were to occur, it would so determine
the action as to bring about a contradiction with itself, which could never
become a universal law.
(LE, 27: 496 (p. 264))
Although it is not stated, it may be presumed that the perfect duty to
‘absolutely speak the truth’ is a perfect duty to others. It is true that, later in
these same lectures, when he contrasts the ethical definition of a lie with the
legal definition of a lie, Kant only talks of a lie violating a perfect duty to
oneself. However, later again in these same lectures Kant acknowledges that
a lie does violate a duty to others – ‘the human duty towards others’ – in
addition to violating a duty to oneself – the duty to ‘humanity in our own
person’. It is important that Kant states here that a lie told to others, even if
it harms no one, and hence, does not violate a duty of right to others not to
lie, nevertheless violates an ethical duty to others not to lie.
There is always a violation in this of humanity in our own person, as well as in
regard to others; we carry on a traffic with words, but not with thoughts, and
so no communication can be thought of here, that would have reality . . . In
sensu juridico [the sense of right], the mendacium [lie] is a falsiloquium dolosum
in praejudicium alterius [harmful falsehood that damages another], but in senso
KANT AND THE DUTY NOT TO LIE
665
ethico [ethical sense] it is already any deliberate untruth. To utter one cannot
be called in any wise permissible. It violates the human duty towards others, as
it does the humanity in our own person, which is hereby depreciated.
(LE, 27: 700–1 (p. 427))
Kant explains how it is that a lie, even a harmless lie, violates two ethical
duties not to lie, in the course of explaining the difference between telling
a lie and breaking a promise. A person who lies to another person
immediately violates a perfect duty to himself not to lie, and mediately
violates a perfect duty to others not to lie.
There is a difference between lying and breach of trust, or mala fides in pactis
servandis [bad faith in keeping a bargain]. He who makes a promise allows the
other to acquire res vel praestatio facti [the doing of a thing] over him, whose
truth the promiser knew as such. But he acts unfaithfully if, because he regrets
this promise, he fails to carry it out. The liar acts contrary to the duty of
humanity in his own person, and mediately against the duty to others; the
contract-breaker acts contrary to the duty towards others, and mediately
against himself.
(LE, 27: 702 (p. 428))
Kant never refers to an ethical duty to others not to lie to others in any of
his published works after the Groundwork. No mention of it is made, for
example, in the extensive discussion of lying in the Doctrine of Virtue, where
Kant is exclusively concerned with the ethical duty to oneself not to lie to
oneself or to others. No mention is made of it in the Right to lie, where Kant
is concerned with the duty of right to humanity not to lie, and in a footnote
refers to the ethical duty to oneself not to lie to oneself or to others. There is
also no mention made of it in the letter to Maria von Herbert, where Kant is
also exclusively concerned with the ethical duty to oneself not to lie to
oneself or to others.18 One explanation of this could be that the ethical duty
to others not to lie to others is covered by the duty of right to humanity not
to lie, since duties of right are ethical duties from the point of view of virtue.
The footnote in the Right to lie then merely indicates that he does not wish
to also talk about the other ethical duty (to oneself) not to lie.19 However, in
the footnote Kant says the duty to oneself not to lie ‘belongs to ethics, but
what is under discussion here is a duty of right’ (RL, 8: 427 n.1 (p. 612
n.1)). The implication is that the duty of right to humanity not to lie is not
an ethical duty. Furthermore, the ethical duty to others not to lie and the
18
We only possess a copy of this letter, as opposed to the letter itself. However, as Zweig points
out, Kant ‘carefully copied and improved on the preliminary draft of his reply to her – unusual
for Kant’ (C, 579). Thus we can be assured that this copy does not differ from the letter that was
sent, and that it is a considered expression of his position.
19
I would like to thank an anonymous referee for making this suggestion.
666
JAMES EDWIN MAHON
duty of right to humanity not to lie do not appear to be equivalent in
content, even if they are equivalent in extension. According to Kant’s
explanation of the duty of right to humanity not to lie, ‘I indeed do no
wrong to him who unjustly compels me to make the statement if I falsify it’
(RL, 8: 426 (p. 612)). Instead, I do wrong to ‘humanity generally’ (RL, 8:
426 (p. 612)). According to the ethical duty to others not to lie, however, it is
always the case that, when I lie to another person, I wrong that person.
Despite the fact that Kant never refers to an ethical duty to others not to
lie to others in any of his published works after the Groundwork, I believe
that Kant did hold that there is an ethical duty to others not to lie to others.
Before explaining what it means to say that this ethical duty is a perfect
duty, I shall defend the claim that the duty to others under discussion in
Section II of the Groundwork is the duty not to lie, as opposed to some other
duty.
A DUTY NOT TO LIE?
Although I have argued that the duty under discussion in Section II of the
Groundwork is the perfect duty to others not to lie to others, commentators
disagree as to which duty to others is at issue here. Some commentators
have said that the duty to others under discussion is ‘the perfect duty to
others to keep promises’.20 However, this is wrong. The obvious example of
a maxim to consider for a perfect duty to others to keep one’s promises
would be a maxim of breaking a promise to another person in order to gain
money, such as ‘when I believe myself to be in need of money, I shall break
my promise to another person to repay the money I have been given’. But
the maxim that Kant considers in the Groundwork is not a maxim of
breaking a promise to another person. It is a maxim of making a promise to
another person that one does not intend to keep. That is to say, it is the
maxim of telling a particular kind of lie to another person. Kant
distinguishes between the breaking of a promise to another person
(a ‘breach of faith’), and a lie to another person, in the lectures on ethics
given in 1784–5: ‘There are also lies whereby the other is cheated. To cheat is
to make a lying promise. Breach of faith is when we promise something
truthfully, but do not have so high a regard for the promise as to keep it.’
(LE, 27: 449 (p. 204)) He also distinguishes between the two in the lectures
on ethics given in 1793–4, as noted above. Thus the duty under discussion in
Section II of the Groundwork is not the perfect duty to others to keep one’s
promises to others.
20
H. B. Acton, Kant’s Moral Philosophy (London: Macmillan, 1970) 22. Many other
commentators have made the same claim. See also, for example, Grounding for the Metaphysics
of Morals, translated by James W. Ellington (Indianapolis: Hackett, 1981), 31 n.14.
KANT AND THE DUTY NOT TO LIE
667
Other commentators have said that the duty under discussion in
Section II of the Groundwork is not an ethical duty to others, but a duty
of right to others not to promise falsely (or lyingly) to others:
The illustration [in the second section of the Groundwork] of perfect duties to
others—the duty not to promise falsely—is a prototype for the duties of justice
which are later elaborated in the Rechtslehre. The action such duties demand
(in this case not making false promises) is relatively well-defined and is owed to
specifiable others; it can be legally enforced, for example, by giving those who
are defrauded by false promisers rights of redress which can be pursued in the
courts.21
It should be noted that if the duty under discussion in Section II of the
Groundwork is a duty of right to others, then the Groundwork does not
provide an example of an ethical duty that is a perfect duty to others.
However, this is wrong also. The duty is a duty about lying as such, and it is
an ethical duty.
First of all, the similarity, bordering on repetition, between the discussion
of the promise that one does not intend to keep, called a ‘lie’, in Section I of
the Groundwork, and the discussion of the lying promise in Section II of the
Groundwork, points to the conclusion that both discussions are about the
same duty. In Section I of the Groundwork, however, the conclusion that
Kant reaches is that one has ‘To be truthful from duty’, and not merely that
one has to not promise falsely (or lyingly).
Secondly, in Section II of the Groundwork Kant contrasts making a lying
promise to another person with assaulting the freedom and property of
others. In the case of assaults on the freedom and property of others, others’
legal rights are transgressed, whereas in the case of making a lying promise
to another person, her legal rights are not transgressed: ‘This conflict with
the principle of other human beings is seen more distinctly if examples of
assaults on the freedom and property of others are brought forward. For
then it is obvious that he who transgresses the rights of human beings
intends to make use of the person of others merely as a means’ (G, 4: 430
(p. 80)). This contrast between making a lying promise to another person,
and assaulting the freedom and property of others, indicates that making a
lying promise is not a violation of a duty of right to others.
This is consistent with what Kant says in the Doctrine of Right, Part I of
The Metaphysics of Morals, where he argues that one does not violate a duty
of right to others when one lies to another person:
This principle of innate freedom already involves the following authorizations . . . to do to others anything that does not in itself diminish what is theirs,
21
Onora O’Neill, ‘Kant’s Virtues’, in How Should One Live? Essays on the Virtues, edited by
Roger Crisp (Oxford: Clarendon Press, 1996).
668
JAMES EDWIN MAHON
so long as they do not want to accept it – such things as merely communicating
his thoughts to them, telling or promising them something, whether what he
says is true and sincere or untrue and insincere (veriloquium aut falsiloquium
[true statement or falsehood]); for it is entirely up to them whether they want
to believe it or not.
(MM, 6: 238 (p. 394))
Finally, in his lectures on ethics in 1784–5, Kant argues that lying is ethically
impermissible by arguing against the possibility of universalizing a maxim of
lying in order to obtain a large estate: ‘For example, lying, in order to obtain
a large estate, becomes impossible to achieve if practised universally, since
everyone knows the aim already. An immoral action, therefore, is one whose
intention abolishes and destroys itself if it is made into a universal rule.’ (LE,
27: 1428 (p. 71)) And in the lectures on ethics in 1793–4, as noted above,
Kant argues that the non-universalizability of the maxim ‘tell an untruth
whenever he could thereby gain a great advantage’ establishes that ‘You
shall absolutely speak the truth, is an imperative of reason’. In both of these
cases the lie in question is not a lying promise, or at least, is not necessarily a
lying promise. The conclusion in each case, however, is that lying is ethically
impermissible and that there is a perfect duty not to lie.
The duty under discussion in Section II of the Groundwork, then, is the
duty to others not to lie to others. The lying promise is merely an example of
a lie told to others. It is now necessary to explain what is meant by saying
that this duty to others is a perfect duty.
PERFECT DUTIES AND IMPERFECT DUTIES
All ethical duties in Kant’s practical philosophy may be divided into perfect
duties and imperfect duties.22 Perfect duties differ from imperfect duties in
at least two important respects. First of all, perfect duties are not
constrained by any other duties—neither by any other perfect duties, nor
by any imperfect duties. For example, the perfect duty to others not to
murder others is constrained neither by the perfect duty to oneself not
to commit suicide, nor by the imperfect duty to others to be beneficent to
others, since it is never the case that one must commit suicide in order not
to murder others, and since one may never murder others in order to be
beneficent to others. However, imperfect duties are constrained both by all
perfect duties and by all other imperfect duties. For example, the imperfect
22
In my ‘Kant on Lies, Candour and Reticence’ (see n.3) I distinguished between duties of
narrow obligation and duties of wide obligation, rather than perfect duties and imperfect duties.
I now believe that although these two distinctions are equivalent in the Groundwork, they are
not equivalent in the Doctrine of Virtue. Everything that I said about narrow and wide duties in
the earlier article should be read as applying to perfect and imperfect duties.
KANT AND THE DUTY NOT TO LIE
669
duty to others to be beneficent to others is constrained by the perfect duty
to others not to murder others, and by the imperfect duty to oneself to
develop one’s natural gifts, since it is ethically impermissible to murder
others in order to be beneficent to others, and since it is ethically
impermissible to be beneficent to others to the extent of never developing
one’s natural gifts. Thus the perfect duty to others not to murder others is
the duty to others not to murder others, whereas the imperfect duty to others
to be beneficent to others is the duty to others to be ethically permissibly
beneficent to others.
The second major difference between perfect and imperfect duties is that
there is no playroom (‘Spielraum’ (MM, 6: 390 (p. 521)), or discretion, incorporated into perfect duties. There is playroom, or discretion, incorporated
into imperfect duties. There are at least two forms of discretion incorporated
into imperfect duties. There is discretion about when (or how much) to act,
and there is discretion about to whom to act. For example, it is possible to
not murder another person all of the time. One can refrain from murdering
another person continuously and perpetually. Therefore the perfect duty to
others not to murder others does not allow for discretion as to when to not
murder another person. One must always not murder another person (or,
one must never murder another person). By contrast, for example, it is not
possible to be beneficent to another person all of the time. One cannot be
beneficent to another person continuously and perpetually. One must
choose between different occasions on which to be beneficent to another
person. Therefore the imperfect duty to others to be beneficent to others
does allow for discretion as to when to be beneficent to others. One must be
beneficent to others some of the time (or, it is not the case that one must be
beneficent to others all of the time).
There is no discretion about to whom to act incorporated into perfect
duties either. It is possible to not murder Conor, not murder Olivia, not
murder Batiste, etc., simultaneously. Therefore the perfect duty to others
not to murder others is the duty not to murder anyone. However, it is not be
possible to be beneficent to Conor, be beneficent to Olivia, be beneficent to
Batiste, etc., simultaneously. One cannot be beneficent to everyone,
simultaneously. One must choose between those who are to be the recipients
of one’s beneficence. Therefore, the imperfect duty to others to be beneficent
to others does allow for discretion as to who will be the recipient of one’s
beneficence. One must be beneficent to some others (or, it is not the case that
one must be beneficent to all others).
In sum, the perfect duty to others not to murder others, correctly stated, is
the duty to others not to murder any other person, ever, whereas the imperfect
duty to others to be beneficent to others, correctly stated, is the duty to
others to be ethically permissibly beneficent to some others some of the time.
The ethical duty to others not to lie to others is a perfect duty. Hence it is
the duty to others not to lie to any other person, ever. It is an exceptionless
negative duty.
670
JAMES EDWIN MAHON
THE ARGUMENT AGAINST LYING TO OTHERS
DRAWN FROM THE UNIVERSAL LAW FORMULA
The argument outlined above that generates the perfect duty to others not to
lie to others is an application of the Universal Law Formula of the
Categorical Imperative: ‘I ought never to act except in such a way that I could
also will that my maxim should become a universal law’ (G, 4: 402 (p. 57)). If a
person acts on a maxim that cannot be made into a universal law, then she is
performing an action for an end that cannot be performed by everyone, for
that end, and succeed. Lying in order to gain money will only succeed so long
as it is not the case that everyone who needs money lies in order to gain
money. The liar, of necessity, applies a double standard: lying in order to
gain money is permissible for her, but not for others.
To apply a double standard is to exempt oneself from a rule that one
requires others to follow. Or, it is to have two rules: one for others, and one
for oneself. According to Kant, any immoral action involves the application
of a double standard:
If we now attend to ourselves in any transgression of a duty, we find that we do
not really will that our maxim should become a universal law, since that is
impossible for us, but that the opposite of our maxim should instead remain a
universal law, only we take the liberty of making an exception to it for
ourselves (or just for this once) to the advantage of our inclination.
(G, 4: 424 (p. 76))
In order for one to gain money by means of lying it is not only necessary
that others who need money do not lie. It is also necessary that those who
are in a position to lend money do not lie that they will lend money (that
they do not, as Kant says, ‘pay me back in like coin’). And it is also
necessary that any intermediaries involved in getting the money to one
(a servant, a bank teller, etc.) do not lie. Since, in principle, anyone else could
be an intermediary, it follows that a person who lies applies the double
standard that lying is permissible for her, but not for anyone else. A person
who lies exempts herself from the rule (do not lie) that she requires everyone
else to follow. Or, she has two rules: one for everyone else (do not lie), and
one for herself (lie).
The basis for applying a double standard, or having one rule for oneself
and another for everyone else, is a presumption of inequality. It is
permissible for one to lie to another person, but not for the other person to
lie to one, because the other person is not equal to one. To lie to another
person, therefore, is to treat the other person as unequal to oneself. In
particular, it is to treat the other person as inferior to oneself. All persons,
however, are equal, in the sense that all persons are equally possessed of
humanity. There can only be a single standard for our actions. Hence lying
to another person is always ethically impermissible. This is what lies behind
KANT AND THE DUTY NOT TO LIE
671
the argument drawn from the Universal Law Formula of the Categorical
Imperative for the perfect duty to others not to lie to others. One owes it to
others to treat them as equals; hence, one owes it to others not to lie to
others. (Such an argument, it is important to note, makes no mention of
harm).
The same argument generates all other duties to others, such as the perfect
duties to others not to steal, kidnap, torture, rape or murder others. For
example, a thief applies a double standard: taking others’ property is
permissible for him, but not for anyone else. Otherwise, others could take
property from him, and he could never possess anything, which is the end of
theft. To steal from another person, then, is to treat her as an unequal, by
taking property from her, and not permitting her to take property from one.
That is what stealing is. Hence it is ethically impermissible to steal from
others. One owes it to others to treat them as equals; hence, one owes it to
others not to steal from others.
There is a further point to note about the argument drawn from the
Universal Law Formula of the Categorical Imperative for the perfect duty
to others not to lie to others. Kant’s claim is that the maxim of obtaining
money by means of making a lying promise to repay money if it is given is
such that it ‘cannot even be thought without contradiction as a universal law
of nature’ (G, 4: 424 (p. 75)). In general, Kant holds that any maxim that
includes lying as a means to some end cannot even be conceived as a
universal law, since it cannot be the case that any end could be achieved by
means of lying if it was practiced universally. It is not the case that lying in
order to achieve some end could be practiced universally, but that this end
would conflict with other ends that the person has. Rather, lying in order to
achieve any end whatsoever could not be practiced universally. In order for
lying to be a means to achieve any end whatsoever, it must not be the case
that lying to achieve this end is practiced universally. Since lying requires a
double standard in order to succeed, it follows that lying necessarily involves
the treatment of other persons as unequals. Hence, lying is necessarily
ethically impermissible.
THE ETHICAL IMPERMISSIBILITY OF LYING
TO AGGRESSORS
The perfect duty to others not to lie to others prohibits lying to any other
person, ever. Hence there is never an occasion on which it is ethically
permissible to lie to another person. As Kant says in the lectures on ethics in
1784–5:
For example: You are not to lie; if this were to rest upon the principle of selflove, it would run: You are not to lie only if it brings harm your way, but if it
profits you, then it is permitted. If it rested on the moral feeling, then anyone
672
JAMES EDWIN MAHON
not possessed of a moral feeling so fine as to produce in him an aversion to
lying would be permitted to lie. Were it to rest upon education or government,
then anyone educated or living under a régime where lying is permitted, would
be at liberty to lie. But if it rests on a principle that resides in the understanding, then the injunction is absolute: You are not to lie, whatever the
circumstances may be.
(LE, 27: 254 (p. 49))
In all of his published writings Kant insists that it is never ethically
permissible to lie to another person. In the lectures on ethics in 1762–4 Kant
says that ‘The importance of the love of truth is so great that one can almost
never make an exception to it’ (LE, 27: 61 (p. 27)). Here he appears to make
a concession to the excusableness, if not the permissibility, of lying to
another person in certain circumstances: ‘If our untruth is in keeping with
out main intent, then it is bad; but if I can avert a truly great evil only by this
means, then . . . etc. Here goodness of heart takes the place of sincerity.’ (LE,
27: 61 (p. 27)) Nevertheless his last statement on the matter in this discussion
appears to be a defence of the absolute prohibition on lying to another
person:
A white lie is often a contradictio in adjecto [contradiction in terms]; like
pretended tipsiness, it is untruth that breaches no obligation, and is thus
properly no lie. Joking lies, if they are not taken to be true, are not immoral.
But if it be that the other is ever meant to believe it, then, even though no harm
is done, it is a lie, since at least there is always deception.
(LE, 27: 62 (p. 28))
There is one passage in his lectures on ethics given in 1784–5, however, in
which Kant appears to permit lying to another person under certain
circumstances:
Now, however, since men are malicious, it is true that we often court danger by
punctilious observance of the truth, and hence has arisen the concept of the
necessary lie, which is a very critical point for the moral philosopher. For
seeing that one may steal, kill or cheat from necessity, the case of the
emergency subverts the whole or morality, since if that is the plea, it rests upon
everyone to judge whether he deems it an emergency or not; and since the
ground here is not determined, as to where emergency arises, the moral rules
are not certain. For example, somebody, who knows that I have money, asks
me: Do you have money at home? If I keep silent, the other concludes that I
do. If I say yes, he takes it away from me; if I say no, I tell a lie; so what am I to
do? So far as I am constrained, by force used against me, to make an
admission, and a wrongful use is made of my statement, and I am unable to
save myself by silence, the lie is a weapon of defence; the declaration extorted,
that is then misused, permits me to defend myself, for whether my admission
or my money is extracted, is all the same. Hence there is no case in which a
KANT AND THE DUTY NOT TO LIE
673
necessary lie should occur, save where the declaration is wrung from me, and I
am also convinced that the other means to make a wrongful use of it.
(LE, 27: 449 (p. 204))
Before commenting on this passage, it should be noted that Kant says a little
later in the same lecture, ‘But the lie is intrinsically a worthless thing,
whether its intentions be good or bad, because it is evil as to its form; it is
still more worthless, however, when it is also evil as to its matter. For by lies
something evil may always result.’ (LE, 27: 449 (p. 205)) Nonetheless, the
passage in which Kant appears to permit lying to aggressors intent upon
immorality should be addressed.
The ‘necessary lie’ to another person that is introduced in this lecture is a
‘very critical point for the moral philosopher’, because if there are no
restrictions as to what is to count as a necessary lie, then everyone will be left
to judge for herself as to when lying to another person is necessary, and the
prohibition against lying to another person will be undermined. Hence
conditions are provided for ethically permissible lying to others. There are
three conditions that must be met. It is only ethically permissible to lie to
another person when: (i) one is unable to remain silent and a declaration is
‘wrung’ (or ‘extorted’) from one; (ii) one is being attacked, or at least being
threatened with violence, by another person; and (iii) one believes that this
other person intends to use the information contained in the declaration to
act immorally.
The first condition that must be met for a ‘necessary lie’, however, is an
impossible condition. It cannot be the case that one is unable to remain
silent and able to either make a truthful declaration or tell a lie. In a
courtroom cross-examination, for example, when one is directed by a judge
to reply to a question, it is false that one is unable to remain silent. One is
able to remain silent. (Of course, one may be penalized for doing so).
Alternatively, if one has been drugged with a truth serum, hypnotized, etc.,
such that one is unable to remain silent, then it is false that one is able to
either make a truthful declaration or tell a lie. One can only make a truthful
declaration. Likewise, if one has been drugged with a ‘lie serum’,
hypnotized, etc., such that one is unable to remain silent, then it is false
that one is able to either make a truthful declaration or tell a lie. One can
only tell a lie. If it is ever true that a person is unable to remain silent, then it
is false that the person is able to either make a truthful declaration or tell a
lie. And if it is ever true that a person is able to either make a truthful
declaration or tell a lie, then it is false that the person is unable to remain
silent.
Kant talks about a declaration being ‘wrung’ from another person. If a
declaration is not the product of drugs, hypnosis, etc., but is a free act, then
it is not ‘wrung’ from another person. Insofar as a declaration is a free act, a
person chooses either to tell a lie or to be truthful. This is the import of
Kant’s favourite quotation from Juvenal referred to earlier: ‘if summoned to
674
JAMES EDWIN MAHON
bear witness in some dubious and uncertain cause, though Phalaris himself
should dictate that you perjure yourself and bring his bull to move you,
count it the greatest of all iniquities to prefer life to honor and to lose, for
the sake of living, all that makes life worth living.’ (CPR, 5: 158–9 (p. 267
n.1)) If one is able to remain silent, or be truthful, rather than perjure oneself, and as a result be murdered, then one is able to remain silent, or be
truthful, rather than lie to an aggressor, and as a result have one’s money
stolen.
Granted this problem with the first condition for a ‘necessary lie’, it
should be abandoned. Note, however, that the remaining second and third
conditions remain extremely restrictive. If one is not under threat of violence
from another person oneself, then lying to another person remains ethically
impermissible. It is only ethically permissible to lie in self-defence. If, for
example, the other person is threatening the lives of others (only), then it is
ethically impermissible to lie. Furthermore, even if one is under threat of
violence oneself, one may not lie to another person unless the information
contained in one’s declaration is going to be used for immoral ends. If, for
example, the other person is threatening one’s life, but the information
contained in one’s truthful declaration is merely, say, personal information,
or common knowledge, and the aggressor simply wants to know this, then it
is ethically impermissible to lie. Even if the first condition is abandoned,
then, it is still only ethically permissible to lie to another person in selfdefence when the information contained in one’s declaration is going to be
used for immoral ends.
Some commentators would appear to defend such an amended argument
for ethically permissible lying: ‘One is to ask whether a universal law of
deceiving to save lives is a possible universal law. It seems clear that it is.’23
The problem with this amended argument for ethically permissible lying,
however, is that even in these circumstances lying is ethically impermissible,
by the argument drawn from the Universal Law Formula of the Categorical
Imperative. A maxim for this action, such as ‘when I believe myself to be
under the threat of violence from another person, who will use the
information of my truthful declaration to act immorally, I shall lie to the
other person in order to defend myself and thwart immorality’, cannot be
made into a universal law. This is because, if there were a universal law
according to which everyone, in self-defence and in order to thwart
23
Barbara Herman, ‘Moral Deliberation and the Derivation of Duties’, in The Practice of Moral
Judgment (Cambridge, Mass.: Harvard University Press, 1993) 152. See also Sally Sedgwick,
‘On Lying and the Role of Content in Kant’s Ethics’, Kant-Studien 82 (1991) 42–62. I take it
that this is also the argument of Thomas E. Hill, Jr., when he says
There are many cases where lies could have saved innocent people from being sent to a
miserable death in a concentration camp. Aware of this and other extraordinary
circumstances, many people could honestly say that their maxim is to tell lies, but only
in cases of that sort.
(Immanuel Kant: Groundwork for the Metaphysics of Morals, 163)
KANT AND THE DUTY NOT TO LIE
675
immorality, were to lie to aggressors, then aggressors, knowing this law,
would not believe anyone whom they threatened. They would ‘laugh at all
such expressions as vain pretenses’. Hence, no one in these circumstances
would ever defend herself and thwart immorality by means of a lie. Lying in
order to defend oneself and thwart immorality would not succeed. Since the
effect (self-defence and thwarting immorality) would not be produced by the
cause (lying), it follows that there cannot be a universal law of lying to
aggressors in order to defend oneself and thwart immorality. Acting on such
a maxim, that is, lying to another person in order to defend oneself and
thwart immorality, is therefore ethically impermissible. As another
commentator has pointed out ‘any maxim along those lines would fail to
pass the test for the categorical imperative. Willing to lie is self-effacing as a
universal maxim; the murderer would not believe such a lie if it was a
universal law to lie for the sake of protecting one’s friends.’24
THE ETHICAL PERMISSIBILITY OF LYING TO LIARS
Some commentators, however, have argued that it is ethically permissible to
lie to another person, by an argument drawn from the Universal Law
Formula of the Categorical Imperative:
These reflections might lead us to believe, then, that Kant was wrong in
thinking that it is never all right to lie. It is permissible to lie to deceivers in
order to counteract the intended results of their deceptions, for the maxim of
lying to a deceiver is universalizable. The deceiver has, so to speak, placed
himself in a morally unprotected position by his own deception. He has
created a situation which universalization cannot reach.25
The argument here is that a maxim of lying to someone whom you believe
is lying to you, in order to counteract the intended result of her lie, is
universalizable. That is, a maxim of action such as ‘When I believe that I am
being lied to by another person I shall lie to the other person, in order to
counteract the intended results of her lie’, can be made into a universal law.
When one lies to another person, one does not believe that the other
person believes that one is lying to her. If one believed that she believed that
one was lying to her, then one would not lie to her, since one would not
believe that one would be able to deceive her, which is the point of lying to
24
Jens Timmermann, ‘The Dutiful Lie: Kantian Approaches to Moral Dilemmas’, Kant und die
Berliner Aufklärung: Akten des IX. Internationalen Kant-Kongresses, edited by Volker von
Gerhardt et al. (Berlin & New York: de Gruyter, 2001) 350. See also Norman Gillespie,
‘Exceptions to the Categorical Imperative’, reprinted in Kant und das Recht der Lüge, edited by
Georg Geismann and Hariolf Oberer (Würzburg: Könighausen and Neumann, 1986) 85–94.
25
Christine M. Korsgaard, ‘The right to lie: Kant on dealing with evil’, reprinted in Creating the
Kingdom of Ends (Cambridge: Cambridge University Press, 1996) 137.
676
JAMES EDWIN MAHON
her. One would believe that it is pointless to lie to her. Hence one lies to
another person only when one does not believe that she believes that one is
lying to her.
If there were a universal law whereby people lied to those whom they
believed were lying to them, and liars knew this law, then liars would always
believe the law to be irrelevant. That is, liars would always believe that this
law fails to apply in the case of their victims. This is because liars would not
believe that their victims believe that they are being lied to. Since liars would
not believe this, and assuming that they would not have any reason to
believe that their victims are (spontaneously) lying to them, they would
believe that their victims are telling the truth. Hence liars would believe what
they are told by their victims.
Someone who believed that she was being lied to by another person, and
who did not let on, would be able to counteract the intended result of
the other person’s lie, by lying in return. Because the liar would believe
what she was told, the victim would counteract the intended results of
the other person’s lie. Since the effect (counteracting the intended results
of lying) would be produced by the cause (lying), it follows that there can be a
universal law of lying to those whom one believes are lying to one in order to
counteract the intended results of their lies. Acting on such a maxim, that is,
lying to someone whom you believe is lying to you in order to counteract the
intended results of her lie, is therefore ethically permissible.
Imagine, for example, that your friend asks you to give her money, telling
you that she will pay you back. You believe that your friend is lying to you,
and that she has no intention of paying you back. Why you believe that she
is lying to you is not important; you simply do. Imagine also that there is a
universal law of lying to those whom you believe are lying to you in order to
counteract the intended results of their lies. And imagine that both you and
your friend are aware of this law. When you lie in return to your friend, and
tell her that you will lend her money the next day, then, assuming that she
has no reason to believe that you are (spontaneously) lying to her, she will
believe you. Your lie will succeed. This is because she would not be lying to
you unless she did not believe that you believe that she is lying to you. Since
she does not believe that you believe that she is lying to you, she will not
believe that you are following the law of lying to those whom you believe are
lying to you in order to counteract the intended results of their lies. Not
having any other reason to believe that you are (spontaneously) lying to her,
she will believe that you are telling the truth. She will believe that you will
lend her money the next day. In lying to her that you will lend her money the
next day you will have counteracted the intended result of her lie—that is,
her gaining your money. So the universal law of lying to those whom you
believe are lying to you in order to counteract the intended results of their
lies works, even when the liars are aware of the universal law.
It may be argued here that if there were a universal law of lying to liars,
then liars would be extremely cautious about lying to others, and might
KANT AND THE DUTY NOT TO LIE
677
26
desist entirely from lying to others. However, even if no one ever lied to
another person, it would still be true that, if anyone were to lie to another
person, it would be ethically permissible to lie to her in return. Hence, if no
lies were ever told, it would still be true that it is ethically permissible to lie
to liars.
The result is that it is ethically permissible to lie to others, by an argument
drawn from the Universal Law Formula of the Categorical Imperative. It is
ethically permissible to lie to others whom you believe are lying to you. If
follows that there is no perfect duty to others not to lie to others, short of
Kant having a different argument.
THE ETHICAL PERMISSIBILITY OF
MURDERING (ETC.) LIARS
The argument drawn from the Universal Law Formula of the Categorical
Imperative for the ethical permissibility of lying to persons whom one
believes are liars succeeds because of the necessary condition of belief on the
part of liars about their victims’ beliefs. In order for a liar to lie to someone,
she must not believe that her victim believes that she is lying to her. This is
true in the case of a victim lying in return to a liar just as much as it is true in
the case of a liar lying to a victim. In the example given above, the victim
must not believe that her lying friend believes that she is lying in return to
her. Otherwise she can’t believe that she can deceive her friend by means of a
lie that she will lend her money.
However, this argument is not merely an argument for the ethical
permissibility of lying to liars. It is also an argument for the ethical
permissibility of doing anything whatsoever to liars. Maxims of action such
as ‘When I believe that I am being lied to by another person I shall murder
the other person, in order to counteract the intended results of her lie’,27 or
‘When I believe that I am being lied to by another person I shall steal from
the other person, in order to counteract the intended results of her lie’, can
also be made into universal laws. Liars would always believe that these
universal laws fail to apply in their cases, since they would not believe that
their victims believe that they are being lied to. The result is that one would
be able to successfully murder liars and steal from liars. Thus murder and
26
I would like to thank an anonymous referee for this objection.
It may be argued here that murder is the killing of a person, against his or her will, whom one
believes is innocent of any wrong-doing whatsoever, and whom one believes is neither a threat
nor a shield of a threat. Since a person whom one believes is lying is not someone whom one
believes is innocent of any wrong-doing, it follows that it is impossible to murder someone
whom one believes is a liar. If the definition of murder is amended, however, to simply the
killing of a person, against his or her will, whom one believes is innocent of certain kinds of
wrong-doing, and whom one believes is neither a threat nor a shield of a threat, then it is
possible to murder someone whom one believes is a liar.
27
678
JAMES EDWIN MAHON
theft are ethically permissible, according to the argument drawn from the
Universal Law Formula of the Categorical Imperative. The ‘morally
unprotected position’ that the liar places herself in ethically permits one
to do anything whatsoever to a liar.
Since this result is a disastrous one for Kant’s ethics, the success of this
argument, drawn from the Universal Law Formula of the Categorical
Imperative, raises a problem for the Universal Law Formula of the
Categorical Imperative as the source of an argument for a perfect duty to
others not to lie to others.
ANARCHIC ENDS
Another problem with the argument against lying drawn from the Universal
Law Formula of the Categorical Imperative is the problem of a person
having a ‘diseased purpose’,28 or an anarchic end, such as the complete
breakdown of trust between all persons (Kant would probably call this a
diabolical end). A maxim such as ‘In order to create a complete breakdown
in trust between all persons, I will lie to another person’, can be made into a
universal law. If everyone lied to others in order to create a complete
breakdown in trust between all persons, then a complete breakdown in trust
in all persons would occur. Since the effect (a complete breakdown in trust
between all persons) would be produced by the cause (lying), it follows that
acting on the maxim, that is, lying in order to create a complete breakdown
in trust between all persons, is ethically permissible.
It may be argued that the anarchic liar needs to rely upon the trust of
others in order to pursue other ends that she has, and that the end of a
complete breakdown in trust between all persons is an end that conflicts
with other ends that she has. Hence, a rational person cannot will to lie to
others in order to create a complete breakdown in trust between all
persons.29 However, this would only establish that there is an imperfect duty
to others not to lie to others. Kant wishes to argue that lying to others
violates a perfect duty to others, and hence, that a maxim of lying to others
in order to create a complete breakdown in trust between all persons ‘cannot
even be thought without contradiction’ as a universal law. Since such a
maxim can be conceived as a universal law, however, it follows, once again,
that the argument drawn from the Universal Law Formula of the
Categorical Imperative fails to generate the requisite exceptionless negative
duty.
28
This expression is used by Christine Korsgaard in ‘Kant’s Formula of Universal Law’,
reprinted in Creating the Kingdom of Ends, 100. The example that follows is my own, however.
29
I would like to thank an anonymous referee for this objection.
KANT AND THE DUTY NOT TO LIE
679
The end need not be anarchic, however. A maxim such as ‘In order to end
the practice of interrogational torture, I will lie to torturers’30 can also be
made into a universal law. If everyone lied to torturers seeking information,
then the practice of interrogational torture would cease. Since the effect
(cessation of the practice of interrogational torture) would be produced by
the cause (lying), it follows that acting on the maxim, that is, lying in order
to end the practice of interrogational torture, is ethically permissible.
Given these failures, it is best to turn to another argument for the perfect
duty to others not to lie to others, drawn from another formula of the
Categorical Imperative.
THE ARGUMENT AGAINST LYING TO OTHERS
DRAWN FROM THE HUMANITY AS AN END
IN ITSELF FORMULA
Another argument for the perfect duty to others not to lie to others can be
drawn from the Humanity as an End in Itself Formula of the Categorical
Imperative: ‘So act that you use humanity, whether in your own person or in
the person of another, always at the same time as an end, never merely as a
means.’ (G, 4: 429 (p. 80)) The argument against lying drawn from this
formula is also to be found in Section II of the Groundwork:
[H]e who has it in mind to make a false promise to others sees at once that he
wants to make use of another human being merely as a means, without the
other at the same time containing in himself the end. For, he whom I want to
use for my purposes by such a promise cannot possibly agree to my way of
behaving toward him, and so himself contain the end of this action.
(G, 4: 430 (p. 80))
According to this argument, with respect to any other-regarding action, in
order for that action to be an ethically permissible action, it must at least be
the case that it is logically possible for the other person to consent to the
action. The vast majority of other-regarding actions are such that it is
logically possible for the other person to consent to them. Giving another
person a gift is an other-regarding action to which it is logically possible for
the recipient to consent. Saving someone from an oncoming lorry is an
other-regarding action to which is it logically possible for the other person
to consent. Some other-regarding actions, however, are such that it is
not logically possible for the other person to consent to them. Theft,
30
By ‘interrogational torture’ is meant torture for the purpose of obtaining information from
the victim. Not all torture is interrogational torture. Terroristic torture, for example, has the
purpose of intimidating people other than the victim. See Henry Shue, ‘Torture’, in Torture: A
Collection, edited by Sanford Levinson (Oxford: Oxford University Press, 2004) 53f.
680
JAMES EDWIN MAHON
kidnapping, torture, rape and murder are examples of such actions. The
victim ‘cannot possibly agree to my way of behaving toward him’. It is of
course possible for a person to consent to losing a possession, to being
falsely imprisoned, to being electrocuted, to having sexual relations, and to
being killed, by others. However the other-regarding actions of theft,
kidnapping, torture, rape and murder incorporate the further element of
being carried out without the consent of the victim. This is because it is an
element of the other-regarding action that the person’s consent is not sought
or that the person refuses consent. It is not logically possible to consent to
an action one element of which is that one’s consent is not sought or that
one refuses consent. It is not possible, for example, to consent to being
kidnapped by others. It is an element of being kidnapped by others that
one’s consent is not sought or that one refuses consent. If one did consent to
being falsely imprisoned, then one would not be consenting to being
kidnapped, but to being falsely imprisoned. Insofar as a person consents,
she is not being kidnapped.
Other-regarding actions to which there is no logical possibility of the
other person consenting cannot be ethically permissible actions. Necessarily,
they are ethically impermissible actions. To act towards another person in a
way that necessarily precludes her consent is to ignore her capacity to
determine her own ends for herself. However, the capacity to determine
one’s own ends for oneself characterizes what Kant calls humanity: ‘The
capacity to set oneself an end – any end whatsoever – is what characterizes
humanity (as distinguished from animality).’ (MM, 6: 392 (p. 522)) Hence to
act towards a person in a way that precludes the possibility of her giving
consent is to treat her as a thing that does not have humanity. Persons,
however, have humanity. When one fails to treat a person as having
humanity, one fails to treat her as person. Failing to treat a person as a
person is ethically impermissible. What makes theft, kidnapping, torture,
rape and murder necessarily ethically impermissible is that they are otherregarding actions in which, by definition, one fails to treat the other person
as a person. This is true regardless of whether the other person is harmed or
benefits from not being treated as a person. A person may, for example,
benefit from being kidnapped by others. This does not alter the fact that in
kidnapping a person one treats her as a non-person, and not as a person,
and it is ethically impermissible to treat a person as a non-person.31
According to this argument, then, when one lies to another person, one
treats her in such a way that she ‘cannot possibly agree to my way of
31
In the film Ruthless People (Touchstone Pictures, 1986; directed by Jim Abrahams and David
Zucker), the character played by Better Midler is kidnapped and held for ransom by a married
couple, but the end result is that her life is improved in various ways (for example, she exercises
and loses weight, something that she has tried and failed to do in the past). To quote from the
film’s publicity: ‘getting abducted may be the best thing that’s happened to their victim in a long
time’. This does not entail that she was not kidnapped, or that it was not ethically impermissible
for the married couple to kidnap her.
KANT AND THE DUTY NOT TO LIE
681
behaving toward’ her. One acts towards her in a way to which it is not
logically possible for her to consent. To do this is not to treat her as a
person. For example, if one is in need of money, one may tell another person
of this, and ask her for money. This provides the person with the
opportunity to freely make it an end of hers to give one money. To do this is
to treat her as a person. To do this is to treat her as a co-operator in one’s
pursuit of money. Alternatively, one may lie to the other person, and tell her
that one wants to borrow money from her. This deprives her of the
opportunity to freely make it an end of hers to give one money, and provides
her with the false opportunity to freely make it an end of hers to loan one
money. To do this is not to treat her as a person. To do this is to treat her as
a tool in one’s pursuit of money. Once again, this remains true regardless of
whether or not the other person benefits from not being treated as a person.
A person may benefit from being lied to. This does not alter the fact that in
lying to another person one treats her as a non-person, and not as a person.
Since, however, it is ethically impermissible to treat another person as a nonperson, and not as a person, it follows that lying to another person is
ethically impermissible.
This argument for a perfect duty to others not to lie to others, drawn from
a different source, is nevertheless open to an objection.
THE IMPOSSIBILITY OF CONSENTING TO BEING
TOLD A PARTICULAR LIE VS. THE POSSIBILITY
OF CONSENTING TO BEING LIED TO WITHOUT
KNOWING WHICH LIE ONE IS TO BE TOLD
It is true that telling a particular lie to another person is an other-regarding
action to which it is not logically possible for the other person to consent.
The person who is lied to ‘cannot possibly agree to my way of behaving
toward’ her, because she is logically precluded from doing so. She is logically
precluded from doing so because, in order for it to be logically possible for
her to consent to being told a particular lie, it must be logically possible for
her consent to be sought. However, it is not logically possible to seek a
person’s consent to be told a particular lie. To seek a person’s consent to be
told something that one believes to be false is to abandon the action of lying
to her. One cannot consistently ask another person, for example, ‘Do you
consent to my lying to you that I will pay you back the money?’ Insofar as
one is seeking her consent, it cannot be her consent to being lied to that one
is seeking. Rather, it is her consent to being told something believed false by
one that one is seeking. Telling a person something believed false by one is a
different action from lying to her, however. One may seek another person’s
consent to be told a fiction, but being told a fiction is not being lied to. If
indeed a person does consent to being told something believed false by one,
she is not consenting to being told a particular lie. Since it is not logically
682
JAMES EDWIN MAHON
possible for a person’s consent to be told a particular lie to be sought, it
follows that it is not logically possible for a person’s consent to be told a
particular lie to be given.
There is, however, a way to understand the claim that it is logically
possible for a person to consent to being lied to by another person,
according to which it is true. Consider, for example, the case of a person
being lied to by her friends about where they are bringing her this evening,
in order to keep a surprise party a secret.
Intuitively, it seems, it makes a great deal of difference if the person you are
lying to has previously agreed to your doing this . . . And if mere hypothetical
consent has force as well, this might also go some way toward justifying
certain ‘white lies.’ For example, we might know that someone would have
agreed to being lied to, for the sake of being thrown a surprise party.32
If this is read as the claim that it is logically possible for a person to consent
to being told a particular lie by her friends, for the sake of being thrown a
surprise party, then the claim is false. If her friend asks her ‘May I lie to you
that we are bringing you to the cinema tonight, for the sake of throwing a
surprise party for you?’, then the action of lying to her has been abandoned
by her friend, and replaced with a different action, that of telling her
something believed false.
However, if this is read as the claim that it is logically possible for a person
to consent to being lied to by her friends, without knowing which lie she is to
be told, for the sake of being thrown a surprise party, then the claim is true. If
her friend asks her, ‘May I lie to you in the future, for the sake of throwing a
surprise party for you?’, then the future act of lying to her has not been
abandoned. It is logically possible to seek a person’s consent to be lied to,
without telling her which lie she is to be told. And it is logically possible for a
person to consent to being lied to, without knowing which lie she is to be told.
That is to say, while it is not possible to consent to being lied to, where ‘being
lied to’ is read de re, i.e. the lie is specified, it is possible to consent to being
lied to, where ‘being lied to’ is read de dicto, i.e. the lie is not specified.
This is only to say, however, that the consent may be sought, and that the
consent may be given. Even if the consent is sought, and the consent is
given, the result of giving such consent may be that no lies are ever told. If
one consents to being lied to by another person, without knowing which lies
one is to be told, for the sake of being thrown a surprise party, then anything
that this person ever says to one afterwards may be a lie, since in principle
the throwing of a surprise party could require lying about absolutely
anything. It follows that one may no longer believe anything that this person
32
Shelly Kagan, Normative Ethics (Boulder, Colorado: Westview Press, 1998) 116.
KANT AND THE DUTY NOT TO LIE
683
33
says, but instead believe that everything that she says is a lie. The other
person will know that one may believe that everything that she says is a lie,
since one’s distrust stems from an agreement that one has made with her.
Hence, she may believe that one believes that everything that she says is a lie.
As was argued above, a person will not lie to another person if she believes
that the other person believes that she is lying to her. If she does believe that
one believes that everything that she says as a lie, then she will not lie to one.
Thus a possible result of one’s agreeing to be lied to by another person,
without knowing which lies one is to be told, for the sake of being thrown a
surprise party, is that the other person will never lie to one.34
However, the other person will know that one may not believe that
everything that she says is a lie. Hence, she may not believe that one believes
that everything that she says is a lie. As a result, she may indeed lie to one,
for the sake of throwing a surprise party for one. Since she has one’s consent
to do so, it follows that she is behaving toward one in a way to which one
has agreed. She has given one the opportunity to freely make it an end of
one’s own to be deceived for the sake of being thrown a surprise party, and
one has freely made it an end of one’s own to be deceived for the sake of
being thrown a surprise party. She has treated one as a co operator in the
deception of oneself for the sake of being thrown a surprise party. Hence, in
lying to one, she is treating one as a person. It is logically possible, therefore,
for a person to lie to another person and to treat the other person as a
person. Since it is logically possible to lie to another person and to treat the
other person as a person, it follows that it is not ethically impermissible to lie
to another person. It is ethically permissible to lie to another person who has
consented to being lied to, without knowing which lies she is to be told. It is
ethically permissible to lie to others by the argument drawn from the
Humanity as an End in Itself Formula of the Categorical Imperative.
It may be argued here that, even if it is logically possible for one to
consent to being lied to by another person, it is ethically impermissible for
one to consent to being lied to by another person (and hence, it is ethically
impermissible for another person to seek one’s consent to be lied to).35 This
is to change from arguing that one ‘cannot possibly’ consent to being lied to
33
This is, I take it, the conclusion of Korsgaard’s rhetorical question: ‘Perhaps what’s being
envisioned is that I simply agree to be lied to, but not about anything in particular. Will I then
trust the person with whom I have made this odd agreement?’ (Korsgaard, ‘The right to lie:
Kant on dealing with evil’, 156 n.5)
34
The implication of Korsgaard’s rhetorical question (see n.33 above) is that if one consents to
be lied to by another person, without knowing which lies one is to be told, then one will believe
that everything that the other person says is a lie. Jonathan Adler has argued, against
Korsgaard, that even if one consents to be lied to by another person, without knowing which
lies one is to be told, nevertheless one may not believe that everything that the other person says
is a lie: ‘Contrary to Korsgaard, the answer to her rhetorically intended question, when applied
realistically, is sometimes ‘yes’.’ (‘Lying, Deceiving, or Falsely Implicating’, Journal of
Philosophy 94 (1997): 441)
35
I would like to thank an anonymous referee for this objection.
684
JAMES EDWIN MAHON
by another person (and hence, that lying is necessarily ethically impermissible), to arguing that one ethically ought not to consent to being lied to by
another person. Consenting to being lied to by another person is consenting
to an attempt to render oneself deceived about some matter, and one
ethically ought not to consent to that, because the attempt to render oneself
deceived about some matter is an action against one’s own humanity. One
would be consenting to the use of humanity in oneself as a mere means for
some other end—in this example, one’s own end of having a surprise party
thrown for one—and humanity in oneself may never be used as a mere means
for some other end, whether it be one’s own end or someone else’s end,
according to the Humanity as an End in Itself Formula of the Categorical
Imperative.
The argument here is that one has a (perfect) duty to oneself not to
consent to being lied to by another person. Notoriously, it is extremely
difficult to demonstrate the existence of duties to oneself, since it requires
that one demonstrate that it is possible to use humanity in oneself as a mere
means for one’s own ends.36 Even if such a duty to oneself could be
demonstrated, however, it would still be the case that the argument for a
perfect duty to others not to lie had failed to show that one ‘cannot possibly’
consent to being lied to by another person, and hence, had failed to show
that lying to another person is necessarily ethically impermissible. It would
still be the case that there is a class of lies to others that this argument is
unable to prohibit, namely, lies told to those who have consented to be lied
to, without knowing which lies they are to be told.
Since the argument drawn from the Humanity as an End in Itself
Formula of the Categorical Imperative fails to generate the perfect duty to
others not to lie to others, it follows that there is no such perfect duty to
others, short of a new argument drawn from some other formula of the
Categorical Imperative. This new argument must be able to prohibit lies told
to those who have consented to being lied to, without knowing which lies
they are to be told. However, I do not believe that any such argument is
forthcoming. Hence there is no perfect duty to others not to lie to others.
CONCLUSION
As was stated earlier, Kant abandoned discussion of a perfect duty to others
not to lie to others in his published works on ethics after the Groundwork in
1785 (assuming, that is, that this was the perfect duty to others under
discussion in the Groundwork). In his later works on ethics he only discussed
the perfect duty to oneself not to lie to oneself or to others. It is doubtful
that he abandoned discussion of the perfect duty to others not to lie to
36
For the best recent attempt to demonstrate the existence of duties to oneself, see Lara Denis,
Moral Self-Regard: Duties to Oneself in Kant’s Moral Theory (NY: Garland Press, 2001).
KANT AND THE DUTY NOT TO LIE
685
others because of the objections considered here. It may have been because
he considered the ethical duty to be redundant, granted that he believed that
all lies to others were prohibited by another ethical duty, the perfect duty to
oneself not to lie to oneself or to others. Or it may have been because he
believed that this ethical could be replaced by a duty of right, namely, the
duty of right to humanity in general not to lie to others, which he argued for
in the Right to lie. In any case, his earlier arguments for this duty, even if
ultimately unsuccessful, remain the most important and powerful arguments
ever made against lying to others.37
Washington and Lee University
37
Earlier versions of parts of this article were read at the Kantian Ethics Conference at the
University of San Diego in January 2003, at the Center for Thomistic Studies at the University
of St. Thomas in February 2004, and at the Joint Session of the Aristotelean Society and the
Mind Association at the University of Kent at Canterbury in July 2004. I would like to thank all
of those who participated in the ensuing discussions, including Tamar Schapiro, Elijah
Millgram, Lawrence M. Hinman, Hans Seigfried, Thomas M. Osborne, R. Edward Houser,
Christopher Martin, Mary Catherine Somers, and Daniel Garber. For discussions about the
arguments contained in this article I would like to thank Onora O’Neill, Andrew J. Cohen, Arif
Ahmed, Alyce Mahon, Mark Shiel, and Joseph Neisser. For comments on a first and second
draft of this article I would like to thank an anonymous referee. For comments on the second
draft I would also like to thank Lara Denis. Work on this article was supported by a Glenn
Grant from Washington and Lee University for the summer of 2003, as well as a pre-tenure
leave in the autumn of 2003. I would like to thank Washington and Lee University for this
generous support. Finally, the first draft of this article was completed while I was a Visiting
Scholar in the Faculty of Philosophy at the University of Cambridge, and a Visiting Fellow of
Clare Hall, Cambridge, in the Michaelmas Term of 2003. I would like to thank the Faculty of
Philosophy, including the secretarial staff, the technical support staff and the library staff, for all
of their help. In particular I would like to thank Simon Blackburn for facilitating my visit. I
would also like to thank the Fellows and staff of Clare Hall for accommodating me, and for
ensuring that my stay was an enjoyable one.