Does Psychological Egoism Entail Ethical Egoism?

JOHN J. TILLEY
INDIANA UNIVERSITY–PURDUE UNIVERSITY INDIANAPOLIS
[email protected]

The Review of Metaphysics 76(1) (2022): 115–133.

Please cite the published version.

I
Psychological egoism is the view, roughly, that every (voluntary) act springs
from the agent’s self-interest. Ethical egoism is the view that if an act is morally
right, this is only because it maximizes the agent’s self-interest. An often
discussed thesis about these views is that psychological egoism, suitably
supplemented with plausible further premises, entails ethical egoism.
Most of those who write about this thesis reject it. However, their rejections
are generally unsatisfying, having at least one of two features.1 The first is a
mere brief treatment of the thesis. This often consists of a short criticism of just
one putative deduction of ethical egoism from psychological egoism, preceded
by little or no discussion of how best to define those two views or how best to
design the deduction. (Such treatments are often unavoidable, especially in
introductory texts that cover much ground.) The second is a formulation of
psychological egoism as the view—call it PE1—that owing to human nature,
nothing ultimately moves us to act but our belief that the act is in our best
interest.
The trouble with a brief treatment is that it is unlikely to be persuasive. It
invites the thought that the thesis examined has been given short shrift. The
trouble with formulating psychological egoism as PE1 is that, as I show later,
PE1 is unfair to psychological egoism.
The thesis that psychological egoism (suitably supplemented) entails ethical
egoism strikes some philosophers as true.2 And from at least one of them, John

1
Illustrations of this point include William K. Frankena, Ethics, 2nd ed. (Englewood Cliffs, NJ:
Prentice-Hall, 1973), 20–23; Fred Feldman, Introductory Ethics (Englewood Cliffs, NJ: Prentice-Hall,
1978), 89–90; Terrance C. McConnell, “The Argument from Psychological Egoism to Ethical Egoism,”
Australasian Journal of Philosophy 56, no. 1 (1978): 41–47; James W. Cornman, Keith Lehrer, and George
Sotiros Pappas, Philosophical Problems and Arguments: An Introduction, 4th ed. (Indianapolis: Hackett,
1992), 303–04; Jordan Howard Sobel, “Egoisms, Psychological and Ethical,” Cogito 10, no. 1 (1996): 22–
28; Robert Almeder, Human Happiness and Morality (Amherst, NY: Prometheus, 2000), 28–29; Louis P.
Pojman and James Fieser, Ethics: Discovering Right and Wrong, 8th ed. (Boston: Cengage, 2017), 78–83;
and Lewis Vaughn, Doing Ethics: Moral Reasoning, Theory, and Contemporary Issues, 5th ed. (New
York: W. W. Norton, 2019), 87–88.
2
See, for example, W. T. Stace, The Concept of Morals (New York: Macmillan, 1937), 209; Steven
Lukes, Individualism (Colchester: ECPR Press, 2006), 87; John Deigh, An Introduction to Ethics
(Cambridge: Cambridge University Press, 2010), chap. 2; and Chris Naticchia, “Hobbesian Realism in
International Relations: A Reappraisal,” in Hobbes Today: Insights for the 21st Century, ed. S. A. Lloyd
(Cambridge: Cambridge University Press, 2013), 241–63, at 250 n. 43.
 p. 2

Deigh, it receives a detailed defense (see section 6). It has further significance
owing to the perennial attraction (to many, at least) of psychological egoism.3
Thus, even if, as I argue in this paper, it ultimately deserves rejection, it warrants
better treatment than it generally receives.
I thus examine several efforts, two of my own devising, to deduce ethical
egoism from psychological egoism (sections 5 and 6). I begin, however, with
important preliminaries. In sections 2 and 3 I deal mainly with properly formu-
lating ethical and psychological egoism. In section 4 I raise the idea that we can
deduce ethical egoism from psychological egoism through premises that include
ought-implies-can, the claim that we can be morally obligated to do an act only
if we are able to do it. I distinguish four broad varieties of such deductions and
identify the only possibly feasible one.

II
Let us start by charitably interpreting the thesis to be examined: that psycholog-
ical egoism, suitably supplemented with plausible further premises, entails
ethical egoism. In particular, let us read the phrase “suitably supplemented with
plausible further premises” to mean “conjoined with certain plausible further
assertions, the resulting set of premises (psychological egoism and the further
assertions) being consistent and nonquestion-begging.”
Next, let me clarify ethical egoism. My earlier statement of it was service-
able, but a fuller, more precise statement is the following:
EE1. An act is morally right if and only if (and because), in terms of
(furthering the agent’s) self-interest, it is at least as good as any of the
agent’s other options.4
Note that EE1 differs significantly from another tempting formulation of
ethical egoism, namely, that an act is morally right if and only if, in terms of
self-interest, it is the agent’s best option. This formulation is unacceptable. It
implies, implausibly, that if a person has just two available actions, each as good
as the other in terms of self-interest, she acts wrongly if she does either of them.

3
A useful source on this subject is Michael Slote, “Egoism and Emotion,” Philosophia 41(2) (2013):
313–35. Slote shows, among other things, that despite well-known objections to psychological egoism, it is
far from dead among psychologists and philosophers.
4
From here on I leave the parenthetical phrases implicit. Philosophers who understand ethical egoism
along the lines of EE1 are abundant. (Indeed, combined with EE2, below, EE1 captures the standard
understanding of ethical egoism among philosophers.) Examples include Milton Snoeyenbos and James
Humber, “Utilitarianism and Business Ethics,” in A Companion to Business Ethics, ed. Robert E. Frederick
(Oxford: Blackwell, 1999), 17–29, at 18; Robert L. Holmes, Basic Moral Philosophy, 4th ed. (Belmont,
CA: Thomson Wadsworth, 2007), 240; Keith Burgess-Jackson, “Taking Egoism Seriously,” Ethical Theory
and Moral Practice 16, no. 3 (2013): 529–42, at 532–33; Douglas W. Portmore, “Introduction,” in The
Oxford Handbook of Consequentialism, ed. Douglas W. Portmore (Oxford: Oxford University Press, 2020),
1–21, at 5; and Russ Shafer-Landau, A Concise Introduction to Ethics (Oxford: Oxford University Press,
2020), 205.
 p. 3

Often, ethical egoism is formulated not as EE1, which concerns moral

rightness, but as a thesis about moral obligation:
EE2. A person morally ought to do an act if and only if, in terms of self-
interest, the act is her best option.5
In EE2, unlike in EE1, the phrase “her best option” is appropriate. To
replace it with “at least as good as any of her other options” would be unchari-
table. Suppose that in terms of self-interest, acts A and B are a person’s two best
options, and each is as good as the other. Then in terms of self-interest, act A is
at least as good as any of the person’s other options. However, ethical egoists,
fairly interpreted, would not say that the person morally ought to do A. They
would say that she morally ought to do either A or B.
I regard EE1 and EE2 as adequate formulations of ethical egoism. Let me
now discuss formulations of psychological egoism.

III
To begin, note that in addition to PE1 (repeated below), there are two other
common understandings of psychological egoism.6
PE1. Nothing ultimately moves a person to act but his belief (operating
with a desire) that the act is in his best interest. He can do the act only
if he has that belief.
PE2. Nothing ultimately moves a person to act but his belief that the act is
(to some extent) in his interest. He can do the act only if he has that
belief.
PE3. People ultimately desire nothing but (real or perceived) personal
benefits, such as their own health, happiness, or survival. They desire
other things only because they see them as direct or indirect means to
(or elements of) a personal benefit.

5
For understandings of ethical egoism of this kind, see, for example, David Copp, “The Ring of Gyges:
Overridingness and the Unity of Reason,” Social Philosophy and Policy 14, no. 1 (1997): 86–106, at 89;
Thomas E. Hill, Human Welfare and Moral Worth: Kantian Perspectives (Oxford: Clarendon, 2002), 139
n. 19; Harry J. Gensler, Ethics: A Contemporary Introduction, 3rd ed. (New York: Routledge, 2018), 237;
Robert Shaver, “Egoism,” in The Stanford Encyclopedia of Philosophy,
https://plato.stanford.edu/archives/spr2019/entries/egoism/, sec. 2; and Andrew Kernohan, Ethical
Reasoning: Theory and Application (Peterborough, ON: Broadview, 2020), 74.
6
Between the two of them, PE2 and PE3 capture the understandings of psychological egoism in many
sources, including Bruce Russell, “On the Relation Between Psychological and Ethical Egoism,”
Philosophical Studies 42, no. 1 (1982): 91–99, at 95–96; Gregory S. Kavka, Hobbesian Moral and Political
Theory (Princeton: Princeton University Press, 1986), 44; Joshua May, “Psychological Egoism,” in Internet
Encyclopedia of Philosophy, http://www.iep.utm.edu/psychego/, sec. 1; Elliott Sober, “Psychological
Egoism,” in The Blackwell Guide to Ethical Theory, 2nd ed., ed. Hugh LaFollette and Ingmar Persson
(Oxford: Blackwell, 2013), 148–68, at 148, 151; Richard Kraut, “Altruism,” in The Stanford Encyclopedia
of Philosophy, https://plato.stanford.edu/archives/spr2018/entries/altruism/, sec. 2.1; Deigh, Introduction to
Ethics, 33; Shaver, “Egoism,” sec. 1; and Vaughn, Doing Ethics, 87–88.
 p. 4

Owing to the phrase “his best interest,” PE1 is unfair to psychological

egoism. It clashes with the uncontroversial fact that people often strive for
immediate or alluring self-benefits without considering whether another
available act would be better for them overall.
Some might try to repair PE1 by adding these words at the end of it:
“except in cases of irrationality or the like.” But this addition is incomplete.
Should we add to it the phrase, “in which case the person may act without seeing
his act as in his interest” or instead add, “in which case he still must see his act
as in his interest to some degree”? If we add the first of those phrases (and revise
the first sentence in PE1 accordingly), PE1 distorts psychological egoism. It
exempts some voluntary acts from the claim that they occur only if the agent
sees them as in his interest. If we add the second phrase, PE1 is a variation of
PE2—a variation with no advantages over PE2. In fact, it has a disadvantage, for
it has controversial elements that PE2 lacks. For instance, it commits us to the
view that if we act without seeing our act as in our best interest, this reflects
some form of irrationality or the like.
Thus, we should simply discard PE1. We should identify psychological
egoism with PE2 or PE3.7 Those formulations are as good as any available.
Before leaving this subject, allow me two more points. First, although PE3
is indeed a good formulation of psychological egoism, from here on I will put it
aside. For it comes with a tacit component: that a person can do an act only if,
regarding that act, he believes something roughly of the form “This act will
bring about x,” where x is something he desires.8 Consequently, PE3 is not
simply about desires but also about actions. Its key point, entailed by its two
components, is that a person can do an act only if he believes something roughly
of the form “This act will bring about x,” where x is a personal benefit or a
perceived means to such a benefit. This point is similar enough to PE2 that what
I argue about PE2 will apply to PE3, mutatis mutandis.
Second, to be charitable to PE2, let us assume that although seeing an act as
in one’s interest often amounts to believing that the act will contribute positively
to one’s interests, sometimes it amounts simply to believing that the act is no
worse, in terms of self-interest, than any other option. Without that assumption,
cases are possible in which PE2 implausibly implies that a person cannot act. An
example: A person knows that each of his available acts (including “doing
7
The point that we should discard PE1 in favor of something better is not original with me. For
instance, Gregory Kavka made it clearly albeit briefly more than three decades ago—and in a notable book:
Kavka, Hobbesian Moral and Political Theory, 38–39. Even so, I make no apologies for emphasizing it. In
the years since Kavka’s book, philosophers have repeatedly used PE1 or a variant of it to represent
psychological egoism. See, for example, Kurt Baier, The Rational and the Moral Order (Chicago: Open
Court, 1995), 159; Jan Narveson, This is Ethical Theory (Chicago: Open Court, 2010), 143; Norman E.
Bowie, Business Ethics in the 21st Century (Dordrecht: Springer, 2013), 32; Almeder, Human Happiness
and Morality, 28; Cornman et al., Philosophical Problems and Arguments, 302; Hill, Human Welfare and
Moral Worth, 141 n. 24; Kernohan, Ethical Reasoning, 74; Pojman and Fieser, Ethics, 78; and Sobel,
“Egoisms, Psychological and Ethical,” 26.
8
See, for instance, May, “Psychological Egoism”; and Sober, “Psychological Egoism.” Of course, the
belief in question need not be explicit or occurrent.
 p. 5

nothing”) is overall harmful to his interests, each act being equally so. As a
result, although he is able to believe that each of the acts is no worse than the
others in terms of self-interest, he is unable to see any of the acts as in his
interest in a more robust sense.

IV
In light of the previous two sections, the way to interpret the thesis at issue—
that psychological egoism (suitably supplemented) entails ethical egoism—is to
read it to say that PE2 (suitably supplemented) entails EE1 or EE2. Although I
cannot cover every possible attempt to establish that thesis, I will critically
examine four attempts that are as good as any I can find.
The first attempt employs ought-implies-can.9 This method raises two
thoughts. First, if our argument is to conclude with EE1, which contains the term
“morally right” rather than “morally ought,” perhaps we should use, not ought-
implies-can itself (OIC), but a variation of it, namely, right-implies-can (RIC).
EE1. An act is morally right if and only if, in terms of self-interest, it is at
least as good as any of the agent’s other options.
OIC. A person morally ought to do an act only if she can do it.
RIC. An act is morally right only if the agent can do it.
Second, if we decide not to use RIC, preferring instead OIC (perhaps
because we find RIC unclear), then perhaps our attempted deduction should
conclude with EE2. That thesis, like OIC, contains the term “morally ought.”
EE2. A person morally ought to do an act if and only if, in terms of self-
interest, the act is her best option.
Viewing my remarks, we see four varieties of the claim that PE2 and ought-
implies-can entail ethical egoism. Examining them will help us identify the most
viable one.
(A) PE2 and OIC (suitably supplemented …) entail EE1.
(B) PE2 and OIC entail EE2.
(C) PE2 and RIC entail EE1.
(D) PE2 and RIC entail EE2.
It may seem that from the thoughts discussed above, (C) and (B) are the
most feasible of these four theses. But in fact we can reject (B), (C), and (D)
straightaway. Let me explain this regarding (C); similar explanations apply to
(B) and (D).

9
It also assumes that in ought-implies-can, the word “can” has the same sense that it has in PE2. This
assumption is virtually ubiquitous in the literature on egoism; hence I grant it in this paper. Likewise for
right-implies-can, which I introduce momentarily.
 p. 6

Contrary to (C), PE2 and RIC (suitably supplemented) do not entail EE1.
They rule out EE1.10 They imply that an act is morally right only if the agent
believes that the act is in her interest. This implication conflicts with the propo-
sition that even if an act is morally right, its agent might not believe that the act
is in her interest. But that proposition follows from the conjunction of EE1 with
the fact, F, that even if, in terms of self-interest, an act is at least as good as any
of a person’s other options, the person might not believe that the act is in her
interest.11 So PE2 and RIC have an implication that conflicts with something to
which EE1 (conjoined with F) commits us. In that sense, PE2 and RIC rule out
EE1. We can thus reject (C). We can reject (B) and (D) for similar reasons.
This leaves (A), the claim that PE2 and OIC entail EE1. I see no way to
dismiss (A) outright. For instance, although PE2 and OIC imply that a person
morally ought to do an act only if she sees the act as in her interest, and EE1
commits us to the claim that even if an act is morally right, its agent might not
see the act as in her interest, this creates no conflict. The implication carried by
PE2 and OIC is consistent with the claim to which EE1 commits us. This would
be false if by calling an act morally right we implied that the act is (or was, at
the time of action) morally obligatory. But we imply no such thing. If a person
has exactly two morally right options, A and B, she would do a morally right act
by doing A. However, what she morally ought to do is not A, but either A or B.

V
Of the four theses just examined, the only feasible one is (A): that PE2 and OIC
(suitably supplemented …) entail EE1. Unfortunately, I can produce no defense
of (A) better than this:
PE2. Nothing ultimately moves a person to act but his belief that the act is
in his interest. He can do the act only if he has that belief.
OIC. A person morally ought to do an act only if he can do it.
1. Thus (from OIC and PE2), a person morally ought to do an act only if
he believes that the act is in his interest.
2. If an act is morally right, then the relevant person (the one for whom
the act is an option) morally ought to do it.

10
Russell and McConnell each make a similar point—though about a thesis more like (B) than (C), and
concerning PE1 rather than PE2. See Russell, “On the Relation Between Psychological and Ethical
Egoism,” 93, 94–95; and McConnell, “The Argument from Psychological Egoism to Ethical Egoism,” 45–
46. Let me add that PE2 and RIC would at once entail and rule out EE1 if their accompanying premises
were inconsistent. We are assuming no such premises.
11
The claim that F is a fact assumes that the word “option” in EE1 (the thesis from which F borrows it)
is not restricted to acts that the agent sees as in her interest. I doubt that any ethical egoist restricts the term
that way. Even psychological egoists need not do so. The word “option” normally includes acts that are
rationally accessible to an agent even if she does not yet have every belief needed to trigger the act.
 p. 7

3. If a person believes that an act is in his interest, then in terms of self-

interest, that act is at least as good as any of his other options.
4. Thus (from 2, 1, and 3), if an act is morally right, then in terms of
self-interest, that act is at least as good as any of the relevant person’s
other options.
5. If, in terms of self-interest, a particular act is at least as good as any of
the relevant person’s other options, then it is morally right.
6. Thus (from 4 and 5), as EE1 says, an act is morally right if and only
if, in terms of self-interest, it is at least as good as any of the relevant
person’s other options.
This argument fails for multiple reasons. An obvious one is that steps 2 and
3 are patently false. Also, I can find no similar arguments that fare any better.
This result is not surprising. PE2 and OIC imply that a person morally
ought to do an act only if he believes that the act is (to some extent) in his
interest. Hence, they connect what a person morally ought to do with what he
believes about the relation of his actions to his interests. And they do so through
the phrase “only if,” which denotes merely a necessary condition. Also, that
necessary condition mentions the act’s being to some extent in the person’s
interest. By contrast, EE1 connects what is morally right of a person to do with
what is true about the relation of his actions to his interests. And it does so
through words that denote a necessary and sufficient condition. Finally, the
latter condition concerns the act’s being maximally in the agent’s interest.
Owing to these facts, the implication carried by PE2 and OIC is logically very
remote from EE1.

VI
In light of the previous two sections, our search for a successful deduction of
ethical egoism from PE2 should probably set aside OIC and RIC. Let us
consider three deductions that do that.
The first deduction is my own creation. I find it the most viable of those I
discuss. But before I present it I must introduce five terms.
The first two terms are “principle of right action” and “compliance-
friendly.” By the first I mean a nontrivial principle of the form “An act is
morally right if and only if it has property P.” Such a principle is compliance-
friendly just in case, for any informed person in any situation in which the
principle recommends (identifies as right) one or more acts, the person is
psychologically able to do at least one of those acts.
The third term, which appears in my clarification of “compliance-friendly,”
is “informed person.” Such a person has accurate, consistent, detailed beliefs
about the morally relevant features of the recommended act(s). By a “morally
relevant feature” (the fourth term) I mean any feature which, according to one or
 p. 8

more well-known normative moral theories (ethical egoism, utilitarianism,

Kantianism, and so forth), contributes to the rightness or wrongness of the act.
For instance, if, in terms of self-interest, an act is at least as good as any of
the agent’s other options, then if that agent is an informed person, she believes
that the act has that property. She also accepts any implications of that belief—
for instance, that the act is to some extent in her interest.12 And because her
beliefs are consistent, she does not hold (even tacitly) that in terms of self-
interest, the act is not at least as good as any of her other options.
The fifth term is “feasible principle of right action.” Such a principle is,
first, a principle of right action, as defined above. Second, it is feasible in this
sense: it is at least as good, at least as viable, as any other principle of right
action of roughly the same species or putatively of that species. For example, if
a principle of right action can reasonably be deemed a form of utilitarianism,
and if, further, it is feasible, then it is not an inferior form of utilitarianism. It is a
form that remains after we have discarded the formulations of utilitarianism that
are less plausible—or less charitable in any way—than they could be. Similarly,
if a principle of right action is putatively a form of ethical egoism, then if it is
feasible it is EE1. This is because EE1 is the most defensible of the principles of
right action (as opposed to principles of obligation) that present themselves as
ethical egoism.
Now for the argument:
PE2. Nothing ultimately moves a person to act but her belief that the act is
in her interest. She can do the act only if she has that belief.
7. If a person can do an act only if she believes that the act is in her
interest, then a feasible principle of right action, FP, is compliance-
friendly only if, for any informed person in any situation in which FP
recommends one or more acts, the person believes, of at least one of
those acts, that the act is in her interest.
8. If for any informed person in any situation in which FP recommends
one or more acts, the person believes, of at least one of those acts, that
the act is in her interest, then FP is either EE1 or a feasible principle
of right action that recommends nothing but acts that always turn out
to be a subset of the acts EE1 recommends.
9. Therefore (from PE2, 7, and 8), FP is compliance-friendly only if FP
is either EE1 or a feasible principle of right action that recommends
nothing but acts that always turn out to be a subset of the acts EE1
recommends.
10. A feasible principle of right action is true (warranted, well-founded,
rationally acceptable) only if it is compliance-friendly.

12
The final paragraph of section 3 is relevant here.
 p. 9

11. There is a true feasible principle of right action.

12. Any feasible principle of right action which, although different from
EE1, recommends nothing but acts that always turn out to be a subset
of the acts EE1 recommends is not true.
13. Therefore (from 9 through 12), EE1 is true: An act is morally right if
and only if, in terms of self-interest, it is at least as good as any of the
agent’s other options.
This valid argument proceeds from PE2 to EE1, employing a cousin of RIC
at step 10. However, it does not succumb to my earlier point that PE2 and RIC
rule out EE1. First, step 10 contains the term “compliance-friendly,” the defini-
tion of which makes 10 implicitly refer, not to people in general, but to informed
persons. Second, the demonstration (in section 4) that PE2 and RIC rule out EE1
depends on the premise that even if one of a person’s available acts is at least as
good, in terms of self-interest, as any of her other options, she might not believe
that the act is in her interest. That premise is not true of informed persons. By
definition, if an informed person has an option which, in terms of self-interest, is
at least as good as any of her other options, then she is aware of that fact, and
thus also aware that the act is to some extent in her interest. Consequently, the
demonstration that PE2 and RIC rule out EE1 is unsound when adapted to the
claim that PE2 and 10 rule out EE1.
But does the argument for 13 succeed? I suspect that at first glance, the
steps that seem the weakest are 8 and 12. However, they are well armored
against counterexamples. To challenge 12 effectively one must find a feasible
principle of right action with the following two features. First, it is true—or at
least we can make a strong case for saying so. Second, although it identifies acts
as morally right on the basis of something other than the property to which EE1
refers, those identified acts invariably have the property to which EE1 refers.
I can think of no principle of right action with both features. Indeed, I find it
difficult to find one with the second feature. If a principle recommends acts on
the basis of something other than the property of being at least as good, in terms
of self-interest, as any of the agent’s other options, then it is hard to see how the
recommended acts could without exception turn out to have that property.
Perhaps we could turn to theistic solution here, claiming that through
rewards and punishments, God ensures that every act in accord with the princi-
ple in question is also at least as good, in terms of the agent’s self-interest, as
any other option available to the agent at the time of the act. This tactic is
unpromising. Its success depends on much more than the plausibility of theism
and of the view that God rewards and punishes people. For it says that God
distributes rewards and punishments in a very particular way.
I will say no more about premise 12, and I will pass over premise 8. Let me
proceed to premise 11, which I consider a vulnerable part of the argument.
Premise 11 says that there is a true, feasible, nontrivial principle of the form
“An act is morally right if and only if it has property P.” The problem for 11 has
 p. 10

two parts. First, 11 cannot stand without argument. For instance, many
philosophers will reject it for its implication that we can derive every true
judgment of right or wrong from a single principle. Unless we see a complete,
strong defense of 11, we cannot count the argument for 13 successful.
A natural reply is that many philosophers—utilitarians, for instance—
accept premise 11; hence, for many, premise 11 is no obstacle to accepting the
argument for 13. But this just brings us to the second part of the problem. The
philosophers who accept 11—I mean those who do so reflectively, through a
supposedly strong defense of 11—do so because they accept a particular princi-
ple of the kind to which 11 refers. Their commitment to 11 is part of their
commitment to a principle of right action that does not leave P unspecified.13
Indeed, it is hard to imagine a feasible defense of 11 that has no commitment,
not even a tacit one, to what property P is or at least to which of the candidates
for P are more likely than others to actually be P. Until we see such a defense,
we should assume that none exists.
We now see the problem for 11. Suppose the only feasible defense(s) of 11
maintains, at least implicitly, either that P is the property of maximizing the
agent’s self-interest or that the latter property is among the most likely candi-
dates for P. Then to accept the argument for 13, which involves accepting 11,
we must accept a defense of 11 which, ultimately, is either a defense of EE1 or
at least a defense more favorable to EE1 than to many other feasible principles
of right action (those outside the set of likely candidates). Such an argument for
13 is unsuccessful. In a broad sense, its premises are question-begging.
Suppose, on the other hand, that we can find a feasible defense of 11 that
neither defends EE1 nor favors a set of principles that includes EE1. Suppose we
accept 11 on the basis of this defense, and then, partly owing to this, accept the
argument for 13. This is to accept the argument for 13 partly owing to an
argument that defends or favors one or more feasible principles that are rivals of
EE1—act utilitarianism, let us suppose. If that argument is weak, it leaves 11
insufficiently plausible. If it is strong, the premises in the argument for 13
cannot collectively be plausible—at least not plausible enough to make the
argument for 13 succeed. For they entail EE1, which conflicts with act
utilitarianism.
Thus, owing to premise 11, the argument for 13 is either not free from
question-begging premises or not free from insufficiently plausible ones. This is
true unless we can find a complete, strong defense of 11 that remains uncom-
mitted about the content of the principle to which 11 refers. Since I lack
confidence that such a defense is possible, I conclude that the argument for 13
falls short.

13
Two contemporary examples are Alan Gewirth, Reason and Morality (Chicago: University of
Chicago Press, 1978); and Peter Singer, Practical Ethics, 3rd ed. (Cambridge: Cambridge University Press,
2011). Although Gewirth’s ethical monism explicitly concerns obligation, he can be interpreted, with no
unfairness, as defending 11. And Singer clearly defends 11. Predictably, the arguments these philosophers
advance for 11 reduce to their arguments for the specific general principles they endorse.
 p. 11

Let us proceed to the other two arguments to be examined. Their source is

John Deigh’s An Introduction to Ethics. To my knowledge, neither has been
examined in print.
Each argument proceeds from what we can charitably treat as a variation of
PE2. For example, the first argument proceeds from the premise that “the motive
of every intentional action is at bottom the desire to promote one’s own
interests.”14 This is a way of saying that we never do an intentional action unless
we believe that it is in our interest, our own interests being what we ultimately
aim to promote. We can thus see the premise as basically PE2. (We can also see
it as basically PE3, but I will not go into that. It affects nothing that follows.)
Deigh calls the first of the two arguments the “primary argument for
[ethical] egoism.” After introducing it, he repeats it in what he considers an
equivalent, clearer form:15
The primary argument can be given a more explicit statement, one that makes
clearer how it proceeds … to its ethical conclusions. On this more explicit
statement, the argument’s basic premiss is the doctrine of psychological egoism
(that the motive of every intentional action is at bottom the desire to promote one’s
own interests). A second premiss is that if the basic motive of every intentional
action is the same, then the end provided by that motive is the only ultimate end of
every intentional action. From these two premisses, it follows that one’s own
interests, which is to say, one’s own happiness, is the only ultimate end of every
intentional action. And since all right actions are intentional actions—a third
premiss—it further follows that the actor’s own happiness is the only ultimate end
of right action. Hence, the highest good for an individual is his or her own
happiness. Hence, from the doctrine of psychological egoism one can, in a few,
simple steps, arrive at the fundamental principle of egoism.16
In this argument Deigh formulates ethical egoism unusually: as concerning
one’s “highest good.” However, perhaps he sees that formulation as equivalent
to a more conventional one.17 In any case, his argument warrants attention. We
can state it thus:
14. Psychological egoism: The motive of every intentional action is at
bottom the desire to promote one’s own interests. In other words, the
basic motive of every intentional action is the same: it is the agent’s
desire to promote her own interests.

14
Deigh, Introduction to Ethics, 34.
15
When he introduces the argument, Deigh says that “intuitively, it draws on the thought that it makes
no sense to prescribe for someone actions that best promote of some end unless he or she could have a
motive to pursue that end.” Deigh, Introduction to Ethics, 34. This seems like a variation of OIC. But
Deigh apparently sees it as inessential to the argument. It does not appear in the quotation below.
16
Deigh, Introduction to Ethics, 34.
17
See Deigh, Introduction to Ethics, chap. 2, especially pages 25–26.
 p. 12

15. If the basic motive of every intentional action is the same, then the
end provided by that motive is the sole ultimate end of every
intentional action.
16. Therefore (from 14 and 15), the agent’s own interests, which is to say,
her own happiness, is the sole ultimate end of every intentional action.
17. Every morally right action is an intentional action.
18. Therefore (from 16 and 17), the agent’s own happiness is the sole
ultimate end of every morally right action.
19. Therefore, the highest good for any individual is her own happiness.
Deigh says that “plainly, the weight of this argument lies with its first
premiss, the doctrine of psychological egoism. So the question to ask is: How
plausible is it [the first premise]?”18 My own question concerns the other
premises. Insofar as they are plausible and the argument is valid, do they entail a
genuine form of ethical egoism?
The answer is no. Proposition 18 admits of two readings:
18.1. The sole ultimate end the having of which makes an action morally
right is the agent’s own happiness. That is, an act is morally right if
and only if it has only the agent’s own happiness as its ultimate end.
18.2. The sole ultimate end of every action that has the properties, whatever
they are, that make an act morally right is the same as that of every
other intentional action: the agent’s own happiness.
Interpreted as 18.1, proposition 18 does not follow from the earlier steps. At
least, it does not do so if we read those steps plausibly, and in a way that makes
14 and 15 entail 16. (Nor does 18.1 entail ethical egoism, but I will discuss this
only in a note.)19 However, if we interpret 18 as 18.2, then although it follows
from the earlier steps, it fails to entail ethical egoism. It does not imply that the
properties it mentions concern only the agent’s own interests. In fact, 18.2
entails no normative ethical principle.
Given these facts about proposition 18, the argument for 19 fails. It
establishes no form of ethical egoism.

18
Deigh, Introduction to Ethics, 34.
19
When we say that an action’s ultimate end is x, we really mean something about the agent. What we
mean about her makes 18.1 reduce to this: An act is morally right if and only if, first, the agent does the act
because she wants happiness and believes that the act is a means to happiness, and second, she wants
happiness for its own sake, not simply as a means to something else. This thesis does not adequately
express ethical egoism. No ethical egoist would accept it. First of all, the phrase “believes that” makes the
thesis implausible. Also, the thesis speaks merely of happiness, not of maximizing happiness or of maxim-
izing self-interest generally. It thus implies that an act can be morally right even if, as a source of happiness
and other personal benefits, it is clearly inferior to the agent’s other options.
 p. 13

Deigh’s next argument appears after the previous one. It is a “new argument
for [ethical] egoism” that substitutes psychological hedonism for the broader
thesis of psychological egoism.20
[The] new argument … proceeds along similar lines as the primary argument but
with a different psychological doctrine as its basic premiss.… While the [new]
argument’s basic premiss, … psychological hedonism [the view “that the motive
of every intentional action is at bottom either a desire for pleasure or an aversion
to pain”], is different from that of the primary argument, its second premiss is the
same [“that if the basic motive of every intentional action is the same, then the end
provided by that motive is the only ultimate end of every intentional action”].…
Thus its first two premisses combine to yield the proposition that the ultimate end
of every intentional action is to experience pleasure or escape from pain. And from
this proposition and the same third premiss as the primary argument’s, which is
that all right actions are intentional actions, it follows that the ultimate end of
every right action is to experience pleasure or escape from pain. At this point,
however, the two arguments diverge.… [The] fourth premiss [in the new
argument] is that an act is right if the actor could not obtain, by some other action,
more of what he is pursuing or less of what he is trying to escape from and if
nothing else besides what he is pursuing or the absence of what he is trying to
escape from is an ultimate end of his intentional actions. The premiss gives a
formal condition on the rightness of an action within a teleological theory. As
such, it represents a requirement of … economic reason. If one is pursuing, as an
ultimate end, something of which one can enjoy more or less, it is not rational to
opt for less when one knows one can get more without additional cost. When this
requirement is combined with the identification of one’s experiencing pleasure or
escaping from pain as the ultimate end of every right action, it follows that the
only ultimate end of every right action is one’s experiencing as much pleasure and
as little pain as one can. In other words, given the hedonistic account of happiness,
one’s own happiness is the only ultimate end of every right action. It is, therefore,
the highest good for an individual. Thus, using a somewhat more complicated
argument than the primary argument, one can base the fundamental principle of
[ethical] egoism on the doctrine of psychological hedonism.21

We can state this new argument as follows:

20. Psychological hedonism: The basic motive of every intentional action
is the same: it is the agent’s desire for pleasure (including the absence
of pain).
21. (Same as 15.) If the basic motive of every intentional action is the
same, then the end provided by that motive is the sole ultimate end of
every intentional action.

20
Deigh, Introduction to Ethics, 37.
21
Deigh, Introduction to Ethics, 37–39.
 p. 14

22. Therefore (from 20 and 21), the sole ultimate end of every intentional
action is to experience pleasure.
23. (Same as 17.) Every morally right action is an intentional action.
24. Therefore (from 22 and 23), the sole ultimate end of every morally
right action is to experience pleasure.
25. An act is morally right if the agent could not obtain, by some other
action, more of what she is pursuing (or less of what she is trying to
escape) and if nothing else besides what she is pursuing is an ultimate
end of her intentional actions.
26. Therefore (from 24 and 25), the sole ultimate end of every morally
right action is the agent’s experiencing as much pleasure as possible.
27. (Same as 18, except for the introductory phrase.) Therefore (from 26),
given a hedonistic account of happiness, the agent’s own happiness is
the sole ultimate end of every morally right action.
28. (Same as 19.) Therefore (from 27), the highest good for any individ-
ual is her own happiness.
Taken literally, premise 25 is absurd. I suspect that Deigh meant “only if”
rather than “if” in 25, but even with that replacement the premise needs
improvement. Let us read it charitably, which, in part, is to read it in light of the
larger passage in which Deigh states it, especially the sentence, “If one is
pursuing, as an ultimate end, something of which one can enjoy more or less, it
is not rational to opt for less when one knows one can get more without
additional cost.”22 Premise 25 now says this:
25.1. If the sole ultimate end of every morally right action is something, S,
that comes in degrees (pleasure, say), then (partly because it is
irrational to opt for less of such a thing if one knows that one could
get more of it at no additional cost) an action is morally right only if
its sole ultimate end is to experience S maximally—that is, to
experience as much of S as possible.
Even with 25.1 in place of 25, the argument for 28 fails. To see this, let us
rewrite steps 24 through 27 so that their meanings (if we interpret them
plausibly) are especially clear.
24.1. The sole ultimate end of every action that has the properties, whatever
they are, that make an act morally right is the same as that of every
other intentional action, namely, the agent’s own experience of
pleasure.

22
Deigh, Introduction to Ethics, 38–39.
 p. 15

25.2. If the sole ultimate end of every action that has the properties,
whatever they are, that make an act morally right is something, S, that
comes in degrees (pleasure, say), then (partly because it is irrational to
opt for less of such a thing if one knows that one could get more of it
at no additional cost) an action has the properties, whatever they are,
that make an act morally right only if its sole ultimate end is to
experience S maximally—that is, to experience as much of S as
possible.
26.1. Therefore (from 24.1 and 25.2), the sole ultimate end of every action
that has the properties, whatever they are, that make an act morally
right is the agent’s experiencing as much pleasure as possible.
27.1. Therefore (from 26.1), given a hedonistic account of happiness, the
agent’s own happiness is the sole ultimate end of every action that has
the properties, whatever they are, that make an act morally right.
28. Therefore (from 27.1), the highest good for any individual is her own
happiness.
In rewriting 24, 25.1, and so forth as shown here, I have not distorted their
meaning. For instance, on a natural reading, 24 is indeed equivalent to 24.1. In
fact, unless 24 is equivalent to 24.1, the earlier steps fail to entail it unless we
read them in a damaging way, a way that makes some of them false or
invalidates the inference to 22.
Unfortunately, to the extent that proposition 28 even approximates EE1 or
EE2 in meaning, 27.1 does not entail it. To put this another way, 27.1 neither is,
nor entails, ethical egoism, adequately formulated. (Likewise for 26.1.) In fact,
27.1 states no conditions for moral rightness. Consequently, the argument for 28
fails.

VII
To conclude: I reject the thesis that psychological egoism, suitably supple-
mented with plausible further premises, entails ethical egoism. I do so owing not
to a cursory treatment of that thesis or to a treatment that uncharitably represents
psychological egoism. I do so having carefully considered some original,
sophisticated attempts to establish the thesis, having first taken care to formulate
ethical egoism and psychological egoism properly.
Of course, I have not considered every imaginable attempt to establish the
thesis I have rejected. However, I have carried the discussion much farther than
others have done, and the attempts I have addressed are as good as any I can
find. We can be confident, I believe, that ethical egoism is not deducible from
psychological egoism.23
23
For helpful discussions, suggestions and the like, I thank Chad Carmichael, David Estlund, Samuel
Kahn, and the anonymous referees.