Ill/lllllllllllwill: # um 12 v

j 1293 108991 08

 

This is to certify that the

thesis entitled

SIDGWICK AND ETHICAL INTUITIONISM

presented by

JUDITH ANDRE

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in PHILOSOPHY

MA Al. 262:2,”

Major professor

Date AUGUST 9, 1979

0-7639

 

 

 

OVERDUE FINFS ARE 25¢ DER DAY
PER ITEM 1

Return to book drop to remove
this checkout from your record.

 

 

 

 

 

 

SIDGWICK AND ETHICAL INTUITIONISM

BY

Judith Andre

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Philosophy

1979

‘

{La/t,“

/ /y/,
A ///

ABSTRACT

SIDGWICK AND ETHICAL INTUITIONISM

BY
Judith Andre

Sidgwick is an ethical intuitionist. His search
for intuitionist grounds results from the following convic-
tions: that all knowledge is either self-evident or derived
from self-evident premises; that moral statements are non-
empirical; that fundamental moral statements must be non-
tautologous; and that any knowledge of self-evident non—
tautologous statements is intuitive.

By "self-evident" he means that such propositions
may be known to be true as soon as they are understood; but
if necessary their truth can also be argued for. His Intui-
tive Method of Verification, for instance, includes several
ways of "assuring ourselves of the self-evidence of what
appears self-evident." And, in practice, for each of the
fundamental principles which he calls "apparently self-
evident" he advances some kind of support.

Sidgwick's appeals to self-evidence are often linked
with appeals to general assent. In such claims that consen-
sus exists, he is sometimes unclear on whose agreement counts,

to what extent, and on what topics. These problems are latent

Judith Andre

within the Common Sense philosophy of Thomas Reid; they
become explicit with Sidgwick because he refuses to sacrifice
plausibility for consistency. Some of these problems lessen
when the propositions which are supported by such appeals
are sorted into epistemological categories, according to
their degree of generality and fundamentality.

The remaining problems can be stated more precisely.
These involve the relationship between what I call second
level propositions (Sidgwick's "apparently self-evident
axioms," such as Benevolence) and third level prepositions
(more specific moral rules, such as truth-telling). For
Sidgwick the third level propositions are knowledge claims,
but not claims to intuitive knowledge. The second level
prepositions, however, are possible cases of intuitive
knowledge.

Richard Rorty distinguishes two kinds of intuition.
In the weak sense, intuitive knowledge refers to prOposi-
tions capable of being known immediately, but which may also
be supported by argument and explication. In the strong
sense, intuitive knowledge is noninferential knowledge for
which there is no decision procedure.

Sidgwick sought ethical foundations which were
intuitive in the weak sense, but his search was unsuccessful.
First, he held that any adequate set of ethical axioms would

specify an end. But he was unable to explicate the difference

Judith Andre

between a rationally chosen end and one which is not.
Therefore, his claim that the only ultimate good is the
pleasure of sentient beings is a claim to intuitive knowledge
only in the strong sense. Secondly, although he claims that
certain axioms are apparently self-evident, he also held

that these fundamental principles must not conflict. Yet

two of his principles, Rational Benevolence and Rational
Egoism, do seem to conflict. They do not conflict if the
universe is such that benevolence is always rewarded; how-
ever, we do not know that this is the case. Since an intui-
tive statement must be capable of being known without sup-
porting considerations, the claim of at least Rational Egoism
or Rational Benevolence to be intuitive in either sense must
fail.

Several contemporary writers argue that egoism is
not rational. If they are right, Benevolence and Egoism are
no longer equal candidates for intuitive knowledge, and
their conflict poses no problem. These writers attempt
explications of "rationally chosen end" which exclude Egoism.
None of the attempts considered here, however, succeeds.

Sidgwick's attempt to find intuitive ethical founda-
tions, although a century old, points to central questions
yet unresolved. His intuitionism is not a claim to have
found such foundations. It is instead an insistence that

such foundations be rationally evaluable.

For my mother, Margaret Anita Andre

ACKNOWLEDGMENTS

I am grateful for the detailed attention given to
this dissertation by my major advisor, Professor Rhoda
Kotzin, and for her advice and support throughout my gradu-
ate program. I am also indebted to Professor Donald Koch
for his willingness to share his knowledge and appreciation
of Sidgwick, and to my other committee members, Professors
Bruce Miller and Craig Standenbaur, for their time and sug-
gestions.

This dissertation was written over a period of
years. Throughout this often difficult time my father,
Louis Andre, and my aunt, Arlene White, patiently encouraged
me. Friends and colleagues stood by me. I would like in
particular to acknowledge the generous friendship of Mari-

anne Glazek and John Valentine. I counted on them.

iii

Chapter

TABLE OF CONTENTS

I. SIDGWICK'S CONCEPT OF MORALITY . . . . .

I.
II.

III.

IV.

V.

Introduction . . . . . z . . . .
Thesis (2): There is (Apparent)
Moral Knowledge . . . . . . .
Thesis (3): No Moral Statement
is Empirical . . . . . . . . .
Thesis (4): Adequate Moral
Axioms Are Non-Tautologous . .
Conclusion . . . . . . . . . . .

II o SELF-EVIDENCE o o o o o o o o o o o o o

I.

II.

III.

IV.

V.

VI.

III. COMMON

I.
II.

III.
IV.
V.
VI.

Introduction . . . . . . . . .
Sidgwick's Theory of Verification
The Three Methods of Reducing
Error . . . . . . . . . . . .
Sidgwick's Intuitionism: Its
Meaning . . . . . . . . . . .
The Intuitive Method of Verifi-
cation Applied . . . . . . . .
Some Problems in Sidgwick's
Concept of Self-Evidence . . .
Conclusion . . . . . . . . . . .

SENSE O O O I O O O O O O O O O 0

Introduction . . . . . . . . .
Problems Within Sidgwick's Use
of "Common Sense" . . . . . .

Reid and Common Sense . . . . . .
A Clarifying Framework . . . . .
Intuition and Common Sense . . .
Conclusion . . . . . . . . . . .

IV. SIDGWICK AND INTUITIONISM . . . . . . .

I.
II.

III.

Introduction . . . . . . . .

Various Meanings of "Intuition"
in Moral Theory . . . . . . .

Intuition and Proposition of the
Third Level . . . . . . . . .

iv

Page

10

21
26

29
29

30
37
47

62
70

72
72
73
97
110
112

115
115

116
123

Chapter

IV. Intuition and Propositions of
the Second Level . . . . . . . .
V. conCIuSion O O I I O O O O C O O O

V. PROBLEMS WITH SIDGWICK'S INTUITIONISM . .

I. Introduction . . . . . . . . . . .
II. Problems in Finding Intuitive
Foundations . . . . . . . . . .
III. Possible Conflicts and the Dualism
of Practical Reason . . . . . .
IV. Attacks on the Self-Evidence of
Egoism . . . . . . . . . . . . .
V. Conclusion . . . . . . . . . . . .

VI. CONCLUSIONS . . . . . . . . . . . . . . .

BIBLIOGRAPHY O O O O O O O O O O I O O O O O O O

Page

137
155

158
158

161
174

184
203

205
207

CHAPTER I

SIDGWICK'S CONCEPT OF MORALITY

I. Introduction

 

"For the supreme rule of aiming at the general
happiness, as I had come to see, must rest on a fundamental

moral intuition, if I was to recognize it as binding at

1

all." In these words, from the Preface to Methods of

 

Ethics, Henry Sidgwick expresses his unique contribution to
philosophical ethics: his attempt to establish utilitarian-
ism on intuitional grounds. Most or all intuitionists before
him had been deontologists, fierce in their rejection of
utilitarianism. Most or all utilitarians had been empiri-
cists, equally fierce in their rejection of intuited knowl-
edge. My intention in this chapter is to trace the connection
between Sidgwick's ethics and his epistemology, to identify
the elements in his ethical theory which lead him to intui-
.tionism (and so, trivially, to his rejection of empiricism).
In the several places where Sidgwick concludes that

an intuition is necessary as the basis of utilitarianism, he

 

1Henry Sidgwick, The Methods of Ethics, 7th ed.
(London: Macmillan & Co., 1907; reprint ed., New York: Dover
Publications Inc., 1966), p. xix.

 

works from varying premises. Most typically he makes this
claim after pointing out that psychological hedonism cannot
establish ethical hedonism.l At one point he argues from
the failure of Mill's argument:

An aggregate of actual desires, each directed

towards a different part of the general happiness,

does not constitute an actual desire for the gener-

al happiness . . . so that there is a gap in the

expressed argument, which can, I think, only be

filled by some such prOposition as . . . the intui-

tion of Rational Benevolence.
I will argue here that it is Sidgwick's conception of an
adequate moral theory which leads him to intuitionism.

Sidgwick believed a number of things about knowledge

in general and moral knowledge in particular. (Throughout
this dissertation, the words "moral" and "ethical" will be
used interchangeably.) Some of his most important beliefs
are listed below:

1. All knowledge is either self-evident or derived

from self-evident premises.

2. There is (apparent) moral knowledge.

3. Moral statements are non-empirical.

4. Adequate moral axioms are not tautologous.

5. Knowledge which is self-evident, non-tautologous,

and non-empirical is intuitive.

 

lIbid., p. 98.
21bid., p. 388.

6. If there is moral knowledge, some moral statements
must be known intuitively.
For convenient reference, these beliefs will be referred to
as Thesis (1), Thesis (2), and so forth.

The major part of this chapter will support the claim
that Sidgwick did in fact hold each of these six beliefs.

The quotation which begins this chapter is evidence that
Sidgwick held Thesis (6). Thesis (5) is the subject of the
fourth chapter of this dissertation and in the present chap-
ter I will simply assume that Thesis (5) represents Sidgwick's
belief.

Thesis (1) he shares with many philosophers, empiri-
cists as well as non-empiricists. It represents a founda-
tionalist view of epistemology, and is certainly Open to
important objection. These objections will not be considered
here. In Chapter IV, however, I will discuss the implica-
tions of Thesis (1) for Sidgwick's intuitionism, and the
difference which its rejection would make.

For the remaining theses--(2), (3), and (4)--I will
give textual support and consider some of the tensions and
apparent inconsistencies in his position. In several cases
these considerations suggest a modified formulation of the
thesis under discussion. In each case the thesis results
from Sidgwick's concept of an adequate moral theory.

Occasionally I will point out that particular theses

account for some significant differences between Sidgwick

and the empiricist utilitarians as well as between Sidgwick
and the "evolutionary" ethicists of his day. (Thesis (3), for
instance--that moral statements are non-empirical--at least
partially explains his rejection of psychological hedonism.)‘

Following this explication I will attempt to recast
this list of theses as an argument, in which Theses (1)
through (5)-—more exactly, these theses as they have been
expanded and reformulated--are premises, and an expanded
version of Thesis (6) is the conclusion. The changes neces-
sary to make the argument valid reveal some interesting
points about Sidgwick's intuitionism.

Finally, I will point out that the conjunction of
these theses is not held by the classical utilitarians Bentham
and Mill, nor by David Hume. It is only in conjunction that
these premises lead to intuitionism. It follows that one
important reason for Sidgwick's epistemological divergence
from other utilitarians--for his non-empiricism--is this dif-
ference concerning the nature of moral theory.

II. Thesis (2): There is
TApparent) Moral Knowledge

 

 

Sidgwick begins the Preface to the First Edition of
Methods of Ethics with "the simple assumption . . . that

there is something under any given circumstances which it is

right or reasonable to do, and . . . this may be known."1

This is a straightforward claim about the possibility of
moral knowledge. Why then the "apparent" in the initial
formulation of this thesis? Sidgwick is, by the end of
Methods of Ethics, skeptical about the possibility of moral
knowledge. His final position is not entirely clear, but he
at least upholds two claims. First, we commonly speak and
think as if moral statements are cognitive (i.e., capable

of being true or false). Secondly, an adequate moral theory

would found moral obligation on true moral statements.~
Moral Statements as Cognitive

Numerous passages in Methods reveal his conviction
that we regard morality as an object of thought. For instance:
From a philos0phical or epistemological point of
view . . . we regard the world of duty and the world
of fact alike as objects of thought and--real or
supposed--know1edge . . .2
The language expressing this attitude is cognitive.
Sidgwick's comparison between the world of duty and
the world of fact forces him to consider some obvious objec-

tions. In replying to these objections he often sounds as

if he is doing more than describing the way we happen to

 

lIbid., p. v.

21bid.

think; he sounds as if he himself believes that all pr0perly
formulated moral statements are either true or false.

One objection points to the fundamental differences
in our methods of determining truth in the two realms (of
duty and of fact). In matters of fact, it might be claimed,
truth depends essentially on correspondence between thought
and fact; in questions of duty, this is not so. He replies
that this difference is less clearcut than it at first appears
to be: correspondence between thought and fact is difficult
to determine.

Though error may lie in want of correspondence
between Thought and Fact, it can only be shown
and ascertained by showing inconsistency between
Thought and Thought, i.e., precisely as error is
shown in the case of our thought about what ought
to be.

A different kind of objection claims that moral propo-
sitions are simply expressions of emotion or of approval.
Sidgwick responds:

It is absurd to say that a mere statement of my
approbation of truth-speaking is properly given

in the proposition 'Truth ought be spoken'; other-
wise the fact of another man's disapprobation
might equally be expressed by saying 'Truth ought
not to be spoken'; and thus we should have two

coexistent facts sSated in two mutually contradic-
tory propositions.

 

lIbid.

21bid., p. 27.

This passage, incidentally, is a refutation of a subjectivist
view of moral language, a view which became prominent in the
mid-twentieth century. Sidgwick's claim that the two propo-
sitions "Truth ought to be spoken" and "Truth ought not to
be spoken" cannot both be true (because they are contradic-
tory) depends upon the assumption that these moral statements
are cognitive.
He offers another argument against the subjectivist

claim.

The peculiar emotion of moral approbation is . . .

inseparably bound up with the conviction . . . that

the conduct approved is 'really' right--i.e., that

it cannot, without error, be disapproved . . .1
He then contrasts this "peculiar emotion" with the emotion
felt when someone decides that it is right to disregard a
rule which one usually obeys (truth-telling, for example).
This repugnance is not a moral sentiment, and is a "feeling
quite different in kind and degree." The conclusion of this
argument is that our feelings are based on a conviction that

we have knowledge.

However, the conclusions reached in Methods of Ethics

 

cast doubt on these common convictions about the possibility
of moral knowledge. Sidgwick finds that two possibly con-
flicting fundamental ethical principles are equally rational,

and that as a result our apparent knowledge of the pair of

 

lIbid., pp. 26-28.

principles is "illusory." The exact sc0pe of the resulting
skepticism is debatable, on two accounts--what Sidgwick does
hold it to be; and what he should hold it to be. (This

issue is discussed at greater length in Chapter V.) Possibly
Sidgwick can retain his initial assumption about the possi-
bility of moral knowledge in spite of his own skepticism,

for there are cases when the two principles do not conflict.
In those cases, perhaps, we know what is right: the action
which both principles permit. The assumption of the Preface,
however, that "there is something under any given circum-
stances which it is right or reasonable to do, and . . . this
may be known," covers all possible circumstances; this assump-
tion fails. In addition it is plausible to conclude that the

skepticism at the end of Methods of Ethics is complete, that

 

Sidgwick no longer holds, or logically can no longer hold,
that we have any moral knowledge whatever.

It is better, therefore, to view that "simple assump-
tion" as one which Sidgwick set out to examine. We may
assume that moral statements are capable of being true or
false, but if moral knowledge is impossible we must abandon
that conviction.

Because of Sidgwick's final skepticism, Thesis (2)
must be reformulated as a conditional claim:

2. If there is moral knowledge, then some moral
statements are known to be true and some moral

statements are known to be false.

Moral Knowledge and Moral Obligation

Thesis (2), like Theses (3) and (4), points to an
important element in Sidgwick's concept of an adequate moral
theory. Sidgwick insists that moral obligation, if it exists
at all, must be founded in rationality. His skepticism is
the more profound for this basic conviction. Although Sidg-
wick calls "ought" and "right" "too elementary to admit of
any formal definition"1 he nevertheless characterizes obliga-
tion in terms of rationality. This is true even in the case
of hypothetical imperatives--of non-moral statements like,
"If you want x, you ought to do y." Sidgwick points out that
here the "ought" implies the unreasonableness of adopting an
end and refusing to adOpt the means indispensable to its
attainment.2 This is also true of moral action:

It can hardly be denied that the recognition of an
end as ultimately reasonable involves the recogni-
tion of an obligation to do such acts as must
conduce to that end.
There are some problems with what Sidgiick means by "ulti-

mately reasonable end"; these are considered in Chapter V.

For the time being it is sufficient to note that Sidgwick

 

lIbid., p. 32.

21bid., p. 37.

3Ibid., p. 36.

’10

characterizes both the end sought by moral action and the
means of reaching that end as subject to rational evaluation.

What has been established so far? From Theses (l)
and (2), treated as premises, a conclusion can be drawn.
Since other conclusions will be drawn later as further theses
are discussed, this will be labeled (C1).

1. All knowledge is either self-evident or derived

from self-evident premises.

2. If there is moral knowledge, then some moral
statements are known to be true and some moral
statements are known to be false.

If any moral statements are known to be true,

they must be either

a) self-evidently true, or

b) derivable from other self-evidently true
statements.

The second alternative listed in this intermediate conclusion

(derivability from other self-evidently true statements) leads

to the discussion of Thesis (3).
III. Thesis (3): No Moral Statement is Empirical

The third thesis in the list is that ethical state—
ments are non-empirical. To the extent that Sidgwick accepts
this, he upholds the independence of morality; moral statements

are not reducible to statements of physical science nor of

11

psychology, nor do moral statements follow from any collection
of statements from these sciences. This independence of
morality must be discussed under two headings: the nature

of statements of obligation and the nature of statements about
good.

In Sidgwick's own words, "the fundamental notion
represented by the word "ought" or "right" . . . [is] essen-
tially different from all notions representing facts of physi-
cal or psychical experience."1 This passage is at first
reading much like the famous passage in which Hume notes that
statements containing "is" are used, fallaciously, to support
statements containing "ought":

This ou ht or ought not, expresses some new relation
or affirmation . . . a reason should be given for
what seems altogether inconceivable, how this new
relation can be a deductionzfrom others which are
entirely different from it.

Sidgwick is more directly addressing the question of
reducibility, Hume that of derivability. Perhaps one rela-
tionship could be denied while the other is upheld, but Sidg-
wick denies both. Sidgwick, in the passage quoted above,

claims an essential difference between statements of obliga-

tion and descriptions of experience.

 

lIbid., p. 11.

2David Hume, A_Treatise of Human Nature, ed. L. A.
Selby-Bigge (Oxford: Clarendon Press, 1888): p. 471.

12

However, there is another instance in which his
position seems less clear. This occurs in a short article
written for a symposium on the question, "Is the Distinction
Between 'Is' and 'Ought' Ultimate and Irreducible?" His

answer is this:

I think it rash to affirm irreducibility but I am
certainly not satisfied with any proposed reduction
proceeding on the lines of scientific thought . . .
i.e., from a psychological or sociological point of
view . . . from this point of view, it seems impossible
to explain or justify the fundamental assumption on
which they all proceed, that some such judgments are
true and others false. . . . One fact cannot be incon-
sistent with another fact; accordingly, regarded from

a psychological or sociological point of view, A's
judgment, e.g., that all gambling is wrong, does not
conflict with B's judgment that some gambling is

right. . . . The reduction, therefore, of duty to
fact, on this line of thought if strictly pursued,
eviscerates ethical thought of its essential import

and interest.1

Now most of this passage concerns the relevance of an idea's
origins to its truth, an issue which will be treated in a
moment. But the passage includes the phrase "it is rash to
affirm irreducibility" and seems to fall short of the position
I have attributed to Sidgwick. His hesitation stems from his
characteristic intellectual caution; he has not, after all,
directly proven that standards of correctness cannot be stated
in empirical terms. The hesitation also stems from his pur-

pose in this discussion. That purpose was critical rather

 

1Henry Sidgwick, "Is the Distinction Between 'Is' and
'Ought' Ultimate and Irreducible?", Proceedings of the Aristo-
telian Society, n.s., l (1892):89.

 

13

than constructive. Although the symposium itself addressed
the question of reducibility in general, Sidgwick took the
opportunity to continue his battle with the psychologism and
evolutionary naturalism of his day. He was concerned, in
other words, to show that a certain position was mistaken
rather than to demonstrate a general proposition about the
nature of moral language. He would of course hesitate to
claim what he had not set out to prove. Not only does irre-
ducibility follow from his position, however; eventually he
granted this in so many words. Lecturing on the ethics of
T. H. Green, Sidgwick said:

I agree as to knowledge of what ought to be being

irreducible to knowledge of what is, has been, and

will be, and I agree in regarding it as fundamentally

important to establish the right relation between

the two.

Just as he holds that ethical statements cannot be
reduced to empirical statements, so he holds they cannot be
derived from empirical statements. This clahm is the subject
matter of Chapter III, Book I of Methods of Ethics. Of the
arguments which he gives to support this claim, two are

particularly interesting. One is his frequent assertion that

no fact about the origin of a belief (moral or non-moral) is

 

lHenry Sidgwick, Lectures on the Ethics of T. H.
Green, Mr. Herbert Spencer, and J. Martineau (New York: Mac-
Millan Company, 1902; reprint. ed., New York: Kraus Reprint
Co., 1968), p. 2.

14

sufficient to tell us whether or not that belief is true.
Another is his assertion that psychological hedonism does

not establish the truth of any ethical theory. Each of these
assertions is an instance of his general belief that moral
statements cannot be derived from non-moral ones--in these

cases, from psychological claims.
The Irrelevance of the Origins of our Beliefs

Sidgwick points out that to claim that people d2
believe that something is the case is not to claim phat they
ggght to believe it is the case. Just as this is true about
factual matters so it is true about matters of morality.
There is a difference between describing beliefs and evaluat-
ing beliefs, in either the empirical or the ethical realms.
The origins of our beliefs in either realm can neither esta-
blish nor disprove their correctness.

This position about beliefs in the moral realm faced
two kinds of opposition. On the one hand, intuitionists
claimed that moral knowledge is innate and therefore certain.
On the other hand, skeptics claimed that showing the psycho-
logical origins of moral convictions proved that they were
nothing but expressions of psychological drives.

He rejects both of these claims, holding instead that
the origin of a belief is a criterion of neither its truth

nor its falsity. The origin of a belief about geometry, for

15

instance, has no bearing on its truth; so those who explain
ethical hedonism in terms of psychological hedonism are in
no way justifying it. Some other means, in ethics as in
geometry, must be found to prove it true or false.

One might object to this comparison, maintaining that
while the fundamental axioms of geometry are uncontroversial,
those of ethics are highly controversial. But this objection
misses Sidgwick's main point, which is that these controver-
sies can only be settled by appeal to a rational standard.
Pointing out the psychological origins of our beliefs is not
such an appeal. The question is not, Why do we have such
beliefs? but, Why should we regard these beliefs as true?
And--more to the point--why should we regard only some of
these beliefs as true, and others as false? As Sidgwick
points out, "all the competing and conflicting moral princi;
ples that men have anywhere assumed must be equally deriva-
tive; yet, since they conflict, they cannot all be true."1
He points out that consideration of the genesis of the moral
convictions leads neither to the establishment of any parti-
cular ethical theory BEE to skepticism about the possibility
of ethical theory. "Surely there is no reason to single

out for distrust the enunciations of the moral faculty, merely

 

lHenry Sidgwick, "The Theory of Evolution in its
Application to Practice," Mind, 0.3. 1 (January, 1876):54.

16

because it is the outcome of a long process of develOpment."l

He goes on to point out that the moral faculty has a shorter
period of develOpment than does our ability to perform the
calculations of higher mathematics. If we have no doubts
about the latter, he asks, why should we about the former?

So much for the skeptics. As for those who claimed
that "originality" is some kind of argument for the truth of
an ethical conviction, Sidgwick's rejection of the position
of Thomas Reid is typical.

The stress laid on [originality as a criterion of

ethical truth] by Reid and other writers is chiefly

due to a psychological assumption now almost exploded;

viz., that the human mind exists at birth in a con-

dition which, though imperfect . . . is at least free

from positive faults: in which therefore, the exercise

of its cognitive faculties, so far as it is capable

of exercising them, must result in truth.
If one rejects the claim that the human mind at birth is less
liable to error than the adult's mind, then the appearance of
some moral belief among young children gives no special sup-
port for its truth.3

One popular attempt, therefore, to derive moral state-

ments from empirical ones has been shown to be inadequate.

 

lIbid.

2Henry Sidgwick, "The Establishment of Ethical First
Principles," Mind, 4 (January, 1879):109.

3Chapter III of this dissertation discusses in more
detail the relationship between Sidgwick and Reid.

17

Another common attempt at the same kind of derivation occurs
when pe0ple argue from psychological facts (or what they take
to be psychological facts) to ethical claims. Frequently

this kind of argument begins with the human desire for pleasure
and argues for a hedonistic concept of good. This, too, Sidg-

wick claims, is an inadequate argument.
Psychological Hedonism

Sidgwick does not accept psychological hedonism, the
belief that "pleasure or absence of pain to myself is always
the actual ultimate end of my action."1 In arguments reminis-
cent of Bishop Butler's, he rejects the claim as confused,
tautologous, or false.2 Of greater interest for this chapter,
however, is the claim which emerges at various points in the
above discussions, that psychological hedonism would not esta-
blish as true gay ethical system--not even ethical hedonism.
Thus one more attempt to derive moral statements from non-
moral ones fails. Since the tenets of psychological hedonism
are probably not empirical, this is not as such supporting
evidence for Thesis (3). Sidgwick would probably agree to a
much stronger thesis: that moral statements are neither

reducible to nor derivable from non-normative statements. His

 

lSidgwick, Methods, p. 41.
21bido ’ pp. 39-560

18

position (described above) on the moral import of psychological
hedonism certainly is consistent with the stronger theses. For
the purposes of this chapter, however, the statement of Thesis
(3) in terms of "empirical statements" will be retained.
Sidgwick does more than reject the claim that psycho-

logical hedonism would establish the truth of ethical hedonism.
He finds psychological hedonism actually incompatible with any
ethical system which bases its precepts on the demands of
reason. Not only would it be "inconceivable that Reason dic-
tates [the pursuit of] . . . any other line of conduct";1 more
importantly, if we accept psychological hedonism

the proposition that he 'ought' to pursue that line

of conduct becomes no less clearly incapable of being

affirmed with any significance. For a psychological

law invariably realised in my conduct does not admit

of being conceived as 'a precept' or 'dictate' of

reason: this latter must be a rule from ghich I am

conscious that it is possible to deviate.
His argument, therefore, is more than the logical point that
"ought" statements are not derivable from "is" statements. In
a kind of conceptual analysis he argues that the very notion
of moral obligation involves a belief in some kind of freedom.

But what kind of freedom? One cannot hold oneself

morally obligated, he writes, unless one is "conscious that

it is possible to deviate" from the rule in question.

 

lSidgwick, Methods, p. 41. (Emphasis added)

2Ibid.

19

It might at first be thought that this is yet another
area where his concept of an adequate moral theory demands
the rejection of empiricism, for freedom under certain defini—
tions is difficult to account for in empirical terms. Freedom
as "free will"--where that means a rejection of determinism--
may pose such problems; but Sidgwick does not claim that this
kind of freedom is a necessary component of ethical theory.
In fact, after presenting arguments for and against determin-
ism, he is unable to declare either side wrong. But he does
think it "possible and useful to show that the ethical impor-
tance of deciding it one way or another is liable to be
exaggerated."1

He supports this conclusion by examining the claim
that "ought" implies "can." He finds two reasonable inter-
pretations of the claim.

1. "I ought to choose to do x" implies "I can do x."
2. "I ought to choose to do x" implies "I can choose

to do x."
Of these two he believes that the acceptance of moral obliga-
tion entails the acceptance of the first, but not of the
second. He also believes that the first is a thesis compatible
with determinism. There are things I can do, and things I

cannot; some things are physically or psychologically impossible,

 

lSidgwick, Methods, p. 66.

20

while others are not. A determinist can agree with this
distinction between what I can and cannot do. For a deter-
minist "'ought' implies 'can'" means that the statement "he
ought to do x" means (among other things) that it is neither
physically nor psychologically impossible for him to do x.

Since one reasonable interpretation of "'ought' implies
'can'" is compatible with determinism, the suspected incom-
patibility between empiricism and Sidgwick's concept of an
adequate moral theory does not arise. It remains true, how-
ever‘ that Sidgwick denies the derivability of obligation
statements from psychological statements.

Thesis (3), then, insofar as it deals with statements
of moral obligation, is clearly Sidgwick's position. There
is more controversy about his position on statements about
moral value--about statements concerning "good." Debate
centers in particular on whether or not he holds good to be
definable in non-moral terms. Because this issue is also
relevant to Thesis (4) ("Adequate moral axioms are non-
tantologous"), it will be postponed until that thesis is
discussed.

Thesis (3), after the discussion of the past few
pages, can be reformulated as follows:

3. Moral statements of obligation are neither reduc-
ible to nor derivable from empirical statements.

It has been suggested but not established that Sidg-

wick holds a stronger version of Thesis (3). Let that be (3'):

21

3'. Moral statements of obligation are neither reducible
to nor derivable from any non-normative statements.

This stronger version leaves open the question of the deriv-
ability of moral statements from statements which are non-
moral but normative. Statements about logical necessity are
particularly interesting candidates here. This possibility
is mentioned again in the last section of this chapter and in
Chapter V.

IV. Thesis (4): Adequate Moral Axioms
Are Non-Tautologpus

 

Sidgwick maintained that fundamental moral principles
cannot be tautologies. He gives some examples of attempts to
base morality on tautologies, and dismisses them as "sham-
axioms."l His reason for the dismissal is that they are not
"important," give no "valuable information," tell us nothing
"practical." "When examined, they are found to affirm no
more than that it is right to do that which is . . . right

to be done."2

And in fact, the examples he gives are of prin-
ciples which do exactly that. He has Plato telling us that
Good is Virtue and Virtue is knowledge of what is Good,
together with action in accordance with that knowledge. He

finds an even more perfect circle in the Stoics: Life

 

1Sidgwick, Methods, pp. 374-78.
2Ibid., p. 375.

22

according to nature is life according to reason; and reason
directs us to live according to nature.

It might well be asked at this point what Sidgwick
means by "tautology." Does he mean, for instance, that moral
axioms cannot be definitions? For all definitions are tau-
tologies. This is a plausible interpretation, particularly
since Sidgwick has carefully noted that "ought" cannot be
defined. It seems clear, however, that what Sidgwick is dis-
missing in this section on tautologies is circular sets of
definitions rather than definitions as such.1 His reason for
the dismissal supports this interpretation. The key word in
his condemnations is "practical." If fundamental moral prin-
ciples cannot give us practical guidance in life, they are
useless. In fact, they cannot be fundamental moral princi-
ples at all.

But this interpretation of what he means by "tautol-
ogy" might be attacked. Schneewind writes, for instance, in
a synopsis of Sidgwick's position, that "to define 'good' as
'pleasure' is self-defeating if what you wish to hold as a
first principle would then be a tautology, and a tautology

"2

cannot be an ethical first principle. In other words,

 

lSidgwick, Methods, pp. 385, 381.

2Encyclopedia of Philosophy, s.v. "Sidgwick," by
J. B. Schneewind.

23

Schneewind treats the rejection of tautology as a reason for
Sidgwick's refusal to define "good."

It is at least open to debate, however, whether Sidg-
wick does refuse to define "good." In relevant passages he
points out, first, that "good" has different connotations than
"pleasure"; then, that it implies a reference to an absolute
standard. "Good" has different connotations than "pleasure";
that is, to call a thing pleasant is not the same as calling
it good. To call a thing good implies that it gives a specific
pleasure, not just any pleasure. He takes examples of non-
moral good: wine, for instance, is not called good simply
because it is pleasant to look at. It must give the other
specific pleasures associated with a good wine. In addition,
it is not enough that just anyone call the wine good (although
anyone may call it "pleasing to me"):

The affirmation of goodness in any object involves

the assumption of a universally valid standard, which,

as we believe, the judgment of persons to whom we

attribute good taste approximately represents.
The last quotation leads to Sidgwick's main point in this dis-
cussion. It is similar to his point about statements of obli-
gation. To call anything good, even in a non-moral sense,
"involves the assumption of a universally valid standard";

and no description of a psychological state can be that. His

main point, then, is not that "good" cannot be defined, but

 

1Sidgwick, Methods, p. 108.

24

that--whether in a moral or non-moral sense--it cannot simply
be defined as pleasure.

Does he or does he not define "good"? G. E. Moore
has written in a famous passage,

'Good,‘ then, is indefinable; and yet, so far as I
know, there is only one ethical writer, Prof. Henry
Sidgwick, who has clearly recognised and states this
fact.

But less notable philosophers have thought that Sidgwick did
define good. James Newell, for example, claims that Sidgwick

"offers a definition of 'good' in terms of pleasurable con-

. 2 . .
soiousness. . . ." Sidi1ck's own words are these:

It would seem then, that if we interpret the notion
'good' in relation to 'desire,‘ we must identify it
not with the actually desired, but rather with the
desirable:--meaning by 'desirable' not necessarily
'what ought to be desired' but what would be desired,
with strength proportioned to the degree of desir-
ability, if it were judged attainable by voluntary
action, supporting the desirer to possess a perfect
forecast, emotional as well as intellectual, of the
state of attainment or fruition.3

Both definists and non-definists could find some support in
this passage. Rather than try to settle the definist issue,
however, let me summarize three important elements in the dis-
cussion so far. First, Sidgwick indisputably calls "ought"

indefinable (because of its simplicity and unanalyzability).

 

1George Edward Moore, Principia Ethica (Cambridge:
University Press, 1903), P. 17.

2James Newell, "The Philosophy of Henry Sidgwick"
(Ph.D. dissertation, University of Maryland, 1975), p. l.

 

3Sidgwick, Methods, pp. 110-111.

25

Secondly, he insists that "good"--in moral and non-moral
uses--means more than pleasure, because of its implicit appeal
to a standard. Thirdly, he rejects tautologies as axioms in
ethical systems.

And this brings us back to the original question:
what does he mean by "tautology" besides a narrowly circular
set of definitions? Whether or not he would call the inter-
pretation of "good" in terms of "what would be desired by an
omniscient being" a definition, it does not seem to be the
case that he rules out all definitions, as such, as ethical
axioms. He does not identify "tautology" with "definition"
because he is not focusing on the truth-functional character-
istics of a tautology, but on its non-informativeness.

It might also be asked what he means by "axiom."
Chapter III will attempt to characterize those principles
which serve as axioms in Sidgwick's moral theory. For the
time being I will simply call them "the most basic of moral
statements." Thesis (4) can be reformulated now, as follows:

4. If any moral statements are known to be true, the
most basic among them must be non-tautologous.
And once again it is Sidgwick's concept of an adequate moral
theory which leads him here. Frank Taylor argues that for
Sidgwick
Ethics aims at both knowledge about morals and
knowledge for conduct. . . . This view of ethics

is more able to do justice to moral experience than
are positivistic, rationalistic, or theological

26

conceptions . . . (Sidgwick's) system is a mixed
system . . . not completely naturalistic orlnon-
naturalistic, definist or nondefinist .

It is this insistence on practical guidance which leads to
Thesis (4), and this fourth thesis is crucial to the argument

which leads him to intuitionism.
V. Conclusion

The argument now reads as follows:

1. All knowledge is either self-evident or derived
from self-evident premises.

2. If there is moral knowledge, then some moral
statements are known to be true and some moral
statements are known to be false.

(Cl) If any moral statements are known to be
true, they must be either self-evidently
true or derived from other self-evident
statements.

3. No moral statement is empirical; i.e., moral state-
ments are neither reducible to nor derivable from
empirical statements.

4. If any moral statements are known to be true, the

most basic among them must be non-tautologous.

 

lFrank Taylor, "The Right and the Good in the Ethics
of Henry Sidgwick" (Ph.D. dissertation, Boston University
Graduate School, 1969), p. l.

27

These four theses, treated as premises, allow the following
conclusion:
(C2) Therefore, if there are any moral state-
ments known to be true, the most basic
among them are non-empirical and non-
tautologous and either
a) self—evident, or
b) derived from other self-evident, non-
empirical statements.
When Thesis (5) is added, a final conclusion may be reached.
5. Self-evident knowledge which is non-tautologous
and non-empirical is intuitive.

(C ) Therefore, if there is moral knowledge,

3
either
a) the most basic moral statements are known
intuitively, or
b) the most basic moral statements are
derived from some other self-evident,
non-empirical statements.
Of the alternatives offered in (C3) Sidgwick pursued the first.
Chapter V will argue that he could have pursued the second.
It remains to be noted that while each of these pre-
mises is necessary for the conclusion (C3), neither Bentham,

Mill, nor Hume holds the entire set. Sidgwick is alone among

major utilitarians in his concept of morality.

28

David Hume, in the famous passage quoted earlier,
makes the same distinction which is expressed in Thesis (3).
However, he denies the possibility of moral knowledge in the
sense expressed in Thesis (2) and thus would not have been
led to a conclusion about the nature of such knowledge.

Mill's position is probably just the reverse of
Hume's. In the equally famous passage from "Utilitarianism"
he implies that moral statements can be derived from empirical
ones (that ethical hedonism can be derived from psychological
hedonism); and whatever one is to make of the phrase "proofs
persuasive to the mind" he is trying to establish the truth
of his ethical position. He is treating moral statements as
cognitive.

Sidgwick's conception of moral statements as practi-
cal, non-empirical, and cognitive is unique among the classical
utilitarians. It is that unique conception which leads to his
other singularity, his intuitionism. The next chapter begins

a close examination of that intuitionism.

CHAPTER II

SELF-EVIDENCE

I. Introduction

 

The last chapter traced the connection between
Sidgwick's conception of an adequate moral theory and his
conviction that such a theory must be founded on at least
one fundamental intuition. This chapter begins an examina-
tion of the intuitionism which Sidgwick found it necessary to
embrace. First it sketches the theory of verification of
which "intuitive verification" is a part; then it examines
the various tests which Sidgwick includes in his "method of
intuitive verification." Next is an examination of the vari-
ous ethical principles which Sidgwick maintains that we know
intuitively. This examination shows that Sidgwick's position
is Open to charges of incoherence. The final section of this
chapter defends Sidgwick against such charges. It is then
evident that to clarify Sidgwick's use of "intuition" several
other large questions must be asked; and so this chapter

points to the following chapters.

29

30

II. Sidgwick's Theoryjof Verification: The
Three Methods of Reducing Error

Terminology and Assumptions

First, a note on terminology. During Sidgwick's life-
time the word "epistemology" was just coming into common
philosophical use. So by the time Philosophy, Its Scope and
Relations is published in 1902, Sidgwick distinguishes "as
epistemology or theory of knowledge," that aspect of philo-
sophy which "considers the fundamental principles of all
departments of systematic thought, but considers them‘with
the special object of examining their validity and evidence."1
He holds that epistemology deals only indirectly with prin-
ciples of inference; its proper subject is the truth of non-
inferential principles, but it is impossible to discuss
those without dealing to some extent with principles of
inference.2

Before "epistemology" became current, "metaphysics"
was used to include what we today call epistemological ques-
tions. That usage is still apparent in Sidgwick's writings.

Earlier in Philosophy, Its Sc0pe and Relations, he writes

that "the discussion of the validity of cognitions attained

 

1Henry Sidgwick, Philosophy, Its Scope and Relations
(New York: MacMillan Co., 1902; New York: Kraus Reprint Co.,
1968), p. 111.

2Ibid., pp. 116-17.

31

otherwise than by inference takes us into Metaphysics."l So
epistemology has two parts: 1) logic, and 2) the truth of
fundamental, underived principles. The second of these parts
is itself part of another subject: metaphysics, which also
includes the non-epistemological question of the nature of
being. Epistemology and metaphysics, in Sidgwick's terms,
intersect; neither is a subset of the other. In this dis-
sertation I will use "epistemology" as Sidgwick does--the
word has the same meaning today--and avoid the word "meta-
physics" except in direct quotes.
Sidgwick holds that certain underlying assumptions

are implicit in any epistemology:

That some beliefs are true and others false, and that

there is some way or ways of distinguishing the one

sort from the other; and [that] the systematic knowl-

edge of these ways is an indispensable element of the

systemizat1on of rational thought.
This systematization he takes to be the primary goal of
philosophy, whose purpose, he writes, is the "exhibition of

"3 These beliefs

system and coherence in a mass of beliefs.
are then more easily purged of error; they are also more easily

grasped and more useful; we are less likely to overlook any

 

lIbid., p. 82.
2Ibid., p. 112.
3Sidgwick, "Criteria of Truth and Error," in Lectures

on the Philosophy of Kant, ed. James Ward (New York: MacMillan
Co., 1905), p. 466.

32

hiatus.1

Sidgwick's methods of verification grow from these
underlying assumptions. They help distinguish truth from
falsity, and do this in part by "exhibiting system and coher-

ence" in our beliefs.
The Three Methods of Verification

To Sidgwick, epistemology begins with the recognition
of error.

To the phiIOSOphical mind the ascertained erroneus-
ness of some beliefs is apt to suggest the possible
erroneousness of all. . . . What guarantees me
against a similar discovery in respect of any other
belief which I am now holding to be true? . . . [I]t
is the removal of this philoSOphical uncertainty-~in
respect of beliefs that, in ordinary thought, are
commonly assumed to be true--that I regard as the
primary aim of Epistemology.

But he assigns to himself a "humbler task" than that
of finding absolute certainty: "anyone who has realized the
slow, prolonged, tortuous process by which the human intellect
has attained such truth as it has now got, will thankfully
accept something less complete."3 Sidgwick seeks instead ways

of reducing the risk of error.4 Reducing, not eliminating--

 

1Ibid., p. 463. This idea of "system," or the
standards of rationality which a method of ethics should meet,
is more fully discussed in Chapter V.

2Ihid.

3Ibid., p. 445.

41b1d., p. 465.

33.

the former can be reached, the latter probably cannot; and
reduction of error is not only within our efforts but well
worth them. Given the comments quoted earlier, then, the
epistemologist's special task is to reduce the risk of error
in fundamental, underived principles.

He calls his methods of reducing error "methods of
verification." In "Criteria" he lists three; in Book III,
Chapter XI of Methods he lists four. The first two listed
in Methods are treated as one in "Criteria." My initial dis-
cussion follows the listing in "Criteria." Following a general
description of each of the three listed there, I return to the
first method and examine it in greater detail. This more
detailed examination includes the divergent listing in Methods.

The ”methods of verification" are these: (1) Intui-
tive, (2) Discursive, (3) Social. Satisfaction of one or all
of these is no guarantee that we have found truth; satisfac-
tion does, however, reduce the risk of error as much as is
possible.

The first method, Intuitive Verification, is, essen-
tially, the careful examination of a supposedly self-evident
proposition. The key words here are "careful examination";
it is not its subject matter which distinguishes the first
method from the other two, for each of them can also be applied
to supposedly self-evident propositions. The point of the

first method, then, is the kind of test which is suggested,

34

rather than the subject of that test. Sidgwick, I think,
obscures this by the name he has chosen. "Intuitive" seems
to indicate that this method alone deals with the apparently
self evident.
What is this "careful examination"? What procedure
is Sidgwick urging? This is a more complicated question than
at first appears; I will return to it after sketching the
other two methods of verification. For now Sidgwick's summary
statement of this method of verification will suffice: "Assure
yourself of the self—evidence of what appears self-evident,
by careful examination."1
Sidgwick's second method of verification is the Dis-

cursive. Even when a proposition passes the test of Intuitive

Verification, "we may be convinced of error through conflict

 

of the judgment thus apparently guaranteed with some other
judgment . . . equally strongly affirmed. . . ."2 The object
of the Discursive Verification is consistency. The method
may be applied to fundamental or derived propositions--al-
though Sidgwick speaks mainly of the fundamental.

It consists in contemplating the belief that appears

intuitively certain along with other beliefs which
may possibly be found to conflict with it.

 

1Ibid., p. 461.

2Ibid.

3Ibid., p. 462.

35

Of course, he adds, "we are always liable to obtain new
beliefs which will conflict with old ones; therefore this
verification is necessarily fallible." Among the three
methods, Discursive Verification is of special and pre-eminent
importance. One reason for its importance is that this method
is the heart of any systematization, and, as noted above,
systematization is for Sidgwick the primary goal of philosophy.
Another reason for the pre-eminence of Discursive Verification
may be that it alone, by the failure of a proposition to meet
its demands, can tell us that one of our beliefs is false
(although it cannot tell us which). This is the only place

in the discussion where Sidgwick claims that a method leads

to certainty, albeit of a limited and negative kind.

The third method of verification which Sidgwick uses
he calls "Social or Oecumenical." Briefly, it consists of
consensus among those who are competent to judge; such con-
sensus makes it legitimate for us provisionally to accept the
object of the consensus. In a world growing increasingly
complex, this method of verification becomes more and more
useful. The primary function of this method is to point out
probable error; it also functions as a heuristic device.

When Sidgwick summarizes his methods, he again dis-
claims

the pretension of establishing absolute truth or
absolute exclusion of error. But if we find that

an intuitive belief appears clear and certain to
ourselves contemplating it, that it is in harmony

36

with our other beliefs relating to the same subject,
and does not conflict with the beliefs of other
persons competent to judge, we have reduced the risk
of error withlregard to it as low as it is possible
to reduce it.

The crucial epistemological questions arise when a
belief passes two of these tests but fails the third. For
instance, when a belief is not self-evident but is commonly
held and is internally consistent; or when inconsistencies
are found in beliefs which are commonly thought to be self-
evident;2 or, finally, when consensus cannot be reached about
competing, and internally consistent, intuitive judgments.3

The second method of verification, which demands that
our beliefs be consistent, is less problematic than the first
two. It can be more easily formulated and, in some form, is
almost universally accepted. The first and third methods are
less clear and much more controversial. The bulk of this
chapter considers the meaning of "intuition" and related
terms, as Sidgwick uses them. The following chapter will

consider the meaning of "consensus" as it occurs in the third

method and what it reveals about Sidgwick's intuitionism.

 

1Ibid., p. 465.

2Sidgwick believes this is the case with Kant's
antinomies.

3Sidgwick, "Criteria," p. 465.

37

III. Sidgwick's Intuitionism: Its Meaning

This section returns to Sidgwick's first method of
verification and the question which was postponed when this
method was first, briefly, considered: Just what does Sidg-
wick mean by the "careful examination" which will assure us
of a statement's "self-evidence"? A related question concerns
possible differences of meaning among the terms which Sidgwick
uses, more or less interchangeably, for "self-evidence."
Finally, this section examines the ethical propositions which
Sidgwick calls self-evident. What does he mean when he so

labels them?
Terminology

There are a number of terms which Sidgwick uses in
somewhat the same sense. These include "self-evident";
"intuition"; "propositions which pass the test of intuitive
verification”; and "first principle." The broadest of these
is "self-evident"; in Sidgwick's terms these are "beliefs that
in ordinary thought are certain [without qualification] . . .
[and whose] certitude may be distinguished as primary or
independent."1

Sidgwick spends very little time discussing the truth

of empirical statements, either particular or universal.

 

1Ibid., p. 437.

38

Usually he reserves "intuitions" for self-evident statements
other than propositions of particular empirical fact; occa-
sionally, however, the word (or some form of it, e.g., "intui-
tive") applies to all self-evident propositions. The first
Method of Verification, for instance, is the "Intuitive"
method, and it applies to particular empirical propositions
as well as to general prOpositions.l Another term is "first
principle." This is a statement which is used to support
others, but which is not in turn supported by others. It
would, if known to be true, have‘to be self-evident. He does
not consider the possibility of a plurality of first princi-
ples, mutually dependent but, as a set, basic. Anything which
counts as a first principle for Sidgwick will be underived.

Now, as noted above, Sidgwick does not consider the
satisfaction of any of his Methods of Verification to be a
sufficient condition of truth. There is some problem, there-
fore, with his use of the term "self-evident," which carries,
more than the others, a connotation of indisputable truth.
Usually, therefore, but not always, Sidgwick prefaces that
phrase with "apparent" or "subjective."

What does he mean by such terms--self-evident, intui-
tion--besides something which seems to be true and for which
supporting premises cannot be given? One thing he clearly

does not mean is that such propositions are "innate." Many

 

lIbid., pp. 442, 452-3.

39

rationalists identified "intuitive" with "innate," although
it was not always clear what they meant. Sidgwick neither
denies nor affirms the possibility of innate ideas; instead,
he regards the question as irrelevant. His reason for this
is his contention (presented in detail in Chapter I, Section
II) that the origin of a belief cannot establish it as true
or as false. Sidgwick therefore is not concerned with the
origins of those beliefs which he calls "intuitive" and cer-

tainly does not identify them as "innate."
Testing Intuitions

In "Criteria," the first "method of verification,"
the "Intuitive Method," is to "assure yourself of the self-
evidence of what appears self-evident, by careful examina-

tion."1

In Methods Sidgwick lists four "conditions, the
complete fulfillment of which would establish a significant
proposition, apparently self-evident, in the highest degree

2 The first of these conditions is:

of certainty attainable."
"The terms of the proposition must be clear and precise." The
second is: ”The self-evidence of the proposition must be
ascertained by careful reflection." The "careful reflection"

and "careful examination" called for in both sources refers

to many different examining procedures. When Sidgwick expands

 

lIbid., p. 461.

2Sidgwick, Methods, p. 338.

40

upon ”careful examination" in "Criteria," he advises us to
look closely at the clarity and distinctness of our ideas.
For that reason I have chosen to treat the first two condi-
tions listed in Methods as parts of the "Intuitive Method of
Verification" listed in "Criteria."

In that article, the "careful examination" takes dif-
ferent forms, in part because of varying objects of the
examination. In universal prOpositions, Sidgwick advises us
to look closely at the clarity and distinctness of the ideas
and at the intuitive certainty of their connections. When
particular empirical judgments are at issue, he advises us
to attempt to purge observation of inference. These sugges-
tions are reminiscent of rationalist and empiricist criteria,1
and in fact appear in the appendix of an article whose inten-
tion is the refutation of Cartesian rationalism and of the
empiricism of John Stuart Mill. His refutation of empiricism
is not directly relevant to this paper since, as shown in
Chapter I, he holds that ethical statements are neither reduc-
ible to nor derivable from empirical statements. Neverthe-
less I will include a brief summary of that attempted refuta-
tion, for two reasons. First, it supplies some important
general information about Sidgwick's epistemology, for he

rejects empiricism on grounds which are independent of his

 

1The word "criteria” is used here only in the sense
in which Sidgwick himself used it: a criterion is a standard
by comparison with which more probably true statements are
distinguished from less probably true statements.

41
concept of morality. This information supplements the
analysis presented in Chapter I. Secondly, Sidgwick's
attempted refutation explains some of the procedures which
he includes under the general title, "Intuitive Verification."
Sidgwick takes empiricism to be the belief that "the
ultimately valid premises of all scientific reasonings are

cognitions of particular facts"1

and attacks this position on
two grounds. First, the validity of general scientific prin-
ciples cannot be established from particular cognitions alone.
He does not argue this point, but takes it as having been
shown by Mill and many others. His second ground for attack
against empiricism is that cognitions of particular facts are
"intimate blends" of observation and inference. Mill himself
points out that

in almost every act of our perceiving faculties,

observation and inference are intimately blended;

what we are said to observe is usually a compound

result of which one-tenth may be observation and

nine-tenths inference.
The inferences involved include implicit comparisons and
classifications of the present with the past; "and the implied

classification may obviously be erroneous either through inac-

curacy of memory or a mistake in the comparative judgment.“

 

lSidgwick, "Criteria," p. 442.

2John Stuart Mill, Lo ic, Book IV, Chapter i, para-
graphs 1, 2; as quoted by Sidgwick, "Criteria," p. 442.

42

Furthermore, even simple observations “cannot be distinctly
thought at all without [this implicit classificationl."l
To claim that experts can distinguish the immediate

cognition from mediated, inferential knowledge is no solution--
for even if it were true, they could not know someone else's
immediate knowledge, only their own. And how do we deal with
the fact that even among the experts there are disagreements?
Consider

. . . the numerous philOSOphical disputes that have

arisen . . . [e.g.] about the impossibility of find-

ing a self in the stream of psychical experience or

. . . as to the consciousness of free will . . .

surely in view of these and other controversies it

would be extraordinarily rash to claim freedom from

error for our cognitions of psychical fact,21et them

he never so rigorously purged of inference.
Sidgwick summarizes his attack on the empiricist assumption
that individual cognitions are true in this way: "The truth
seems to be that the indubitable certainty of the judgment 'I
am conscious' has been rather hastily extended by Empiricists
to judgments affirming that my present consciousness is such
and such." Empiricism fails because it requires a sharp dis-
tinction (between the inferred and the non-inferred) which it
cannot provide.

One way, therefore, of performing the First Method

of Verification is to attempt to purge our observations of

 

1

2

Sidgwick, "Criteria," p. 453.

Ibid.

43

inference. As the above discussion makes clear, this attempt
makes it more likely that our particular beliefs--beliefs
grounded on particular observations--will be true. It does
not assure us that they will be true.

Just as a rejection of empiricism underlies Sidg-
wick's approach to verifying particular empirical beliefs, so
a rejection of Cartesian rationalism underlies his approach

to universal beliefs.l

Descartes had claimed that any propo-
sition which is clear and distinct is true; but Sidgwick takes
Descartes to task for not satisfying his own requirements--
for failing to make "clear what he means by the 'clearness'

of a notion."2

"Distinct," he thinks, can be more helpfully
explicated; an idea is distinct if it ”is not liable to be
confounded with that of any different object."3 Nevertheless,
this method, too, is only occasionally of any real use.

(Sidgwick does find it applicable retrospectively: i.e., the

 

discovery of error involves the discovery of past confusion.

But "before the discovery of the error the thought eppeared

to be quite free from confusion . . ."4)

 

 

1The reader may wonder about Sidgwick's position on
particular non-empirical beliefs. In the case of particular
ethical beliefs ("Robbing this bank today is wrong") he simply
holds that they cannot be self-evident. See his rejection of
"Perceptional Intuitionism," Methods, pp. 98-100.

2

Sidgwick, "Criteria, p. 449.

3Ibid.

4Ibid., p. 450.

44

Again, then, a kind of "careful examination" is
suggested (in explanation of Intuitive Verification) which is
a more modest version of suggestions made by a major philo-
sopher or philosophical system. Sidgwick rejects empiricism;
he also finds himself compelled to reject the claim of Ration-
alism

to put forward a simple infallible criterion for the
kind of knowledge which is to be taken as the ulti-
mately valid basis of all . . . [other] knowledge.

I regard both criteria as useful, as a means of
guarding against error, but neither as infallible.

l

A third kind of "careful examination" appears when
Sidgwick considers a criterion formulated by Herbert Spencer,
among others. If the negation of a proposition is inconceiv-
able, then the proposition is self-evidently true. Sidgwick
scrutinizes "inconceivable," and concludes that in statements
of particular fact, it means "incredible." I cannot believe

that I am not now thirsty; but I can certainly conceive of

what it would be like. For universal propositions,

 

lSidgwick, "Criteria," p. 452. On the same page he

notes
"a curious interchange of rales between Rationalism
and Empiricism as regards the evidence claimed for
their respective criteria. While the Rationalist's
criterion is partly supported . . . on an appeal to
experience, the validity of the Empirical criterion
appears to be treated as self-evident."
Descartes, for instance, having found one indubitable proposi-
tion--cogito ergo sum-~tried to isolate the characteristics
of that idea which gave it its self-evidence. He finds that
the idea is 'clear and distinct' and concludes from this ex e-
rience of indubitability that he has found a criterion of self-
ev1dence. And Mill's statement that "'whatever was a matter
of direct perception . . . would have been real' seems to have
been regarded as not requiring proof." (p. 452).

45

"inconceivable" may mean "unimaginable"; e.g., "I cannot
imagine or conceive two straight lines enclosing a space."
But "imagine" should not be limited to sensory experience;
what Sidgwick means is broader: to be able to conceive of
something being true is to know what it would be like for it
to be true.

Just as in his treatment of the first two kinds of
"careful examination," here too Sidgwick adds a caution: "my
inability to conceive or believe [may] be due to a defect of

my intellect."l

And, he notes, the history of thought amply
shows us the fallibility of this criterion, too.

The discussion in Methods supplies still more kinds
of "careful examination." The supposedly self-evident propo-
sition must be distinguished from "mere impressions or im-
pulses," from opinions which seem unquestionable because they
have been repeated so often, from claims whose major support
is strong sentiment.2 Making these distinctions will help us
narrow the field to pr0positions whose only support is that.
of reason; at that point we can turn to the methods presented
in "Criteria."

Remember that Sidgwick's second and third Methods of
Verification can also be used to reduce error in apparently

self-evident principles. What we have, then, in the two dis-

cussions mentioned thus far, is at least six different methods

 

11hid., p. 449.

2Sidgwick, Methods, p. 339.

46

of verification. Four of these--purge observation of
inference; formulate the proposition clearly and distinctly;
ask whether its negation is conceivable; separate claims of
reason from claims of sentiment, authority, and tradition--
these four are lumped together to form the First or Intuitive
Method because, unlike the Second and Third Methods, these
are of use only in inspecting apparently self-evident beliefs.
There is yet another course, however, which reveals
Sidgwick's position on the examination of such beliefs. This
is an article entitled, "The Establishment of Ethical First
h

Principles."

[Wle are logically led to a conclusion which yet when

reached we regard as a first principle. We start with

a proposition which appears self-evident; we reflect

on it and analyze it into a more general proposition

with a limitation; . . . we see that [the limitation]

is arbitrary and w/o foundation in reason; we deny its

validity and we substitute for our original principle

the wider statement of which that affirmed a part.
Sidgwick's example involves the principle, "all pain of human
or rational beings is to be avoided." He sees that the limita-
tion to "human or rational beings" is one which reason cannot
support, and is then led to the wider principle, "all pain is
to be avoided."2

This removal of unreasonable limitations from appa-

rently self-evident propositions might reasonably be called

 

lHenry Sidgwick, "The Establishment of Ethical First
Principles," Mind, 4 (January, 1870):107.

21bid., p. 106.

47

still another Method of Verification, one which belongs with
those lumped together into the first, Intuitive Method.

The remainder of this chapter deals with the princi-
ples which Sidgwick calls self-evident, and the discussion
with which he supports that claim. Looking at those princi-
ples will mean looking at Sidgwick's BEE of the first Method
of Verification. By the end of the chapter we will have seen
in detail what Sidgwick means by self-evident. His theoreti-
cal discussion of the concept has been presented in Sections
II and III; his practical application of the concept is pre-
sented in the following pages, Section IV.

IV. The Intuitive Method of
Verification Applied

 

 

Commentators differ markedly in the number of proposi-
tions which they believe Sidgwick held to be self-evident.
The reason for this is that Sidgwick formulates the self-
evident pr0positions in a number of different ways. Schnee-
wind compiles all of these formulations and groups them in
three categories: formulations of the axiom of Justice,
formulations of the axiom of Prudence, and formulations of
the axiom of Benevolence.l The justification for this group-

ing is Sidgwick's frequent references to those three axioms.2

 

1J. B. Schneewind, Sidgwick's Ethics and Victorian
Moral PhilOSOphy (Oxford: The Clarendon Press, 1977), pp. 290-
297.

 

2See for instance Chapter XIII, "Philosophical Intui-
tionism" of Book III in Methods; in particular, p. 382.

48

Sidgwick, however, sometimes speaks in addition of "Rational
Egoism," and Schneewind's justification for calling this a
formulation of Prudence is the close similarity in the content
of the two. As the following discussion will show, I think
Prudence has to be treated as two separate axioms. Further-
more, I do not treat rational egoism as just a formulation of
Prudence, but as a separate principle. It is also not clear
that everything Schneewind has treated as a formulation of the
axiom of Justice deserves that label.

The first two of Sidgwick's formulations which Schnee-
wind lists as formulations of "Justice" are:

1. "Whatever action any of us judges to be right for
himself, he implicitly judges to be right for all
similar persons in similar circumstances."1

2. "If I judge any action to be right for myself, I
implicitly judge it to be right for any other
person whose nature and circumstances do not
differ from my own in some important respect."2
The differences between the two are not important. To

the twentieth century reader these sound very much like prin-
ciples of universalizability. Sidgwick bases them directly

on the meaning of the word "ought."3

 

lSidgwick, Methods, p. 379.

2Ibid., p. 209.

3They are not, of course, definitions. "Ought" can-
not be defined; "it can only be made clearer by determining

49

Does Sidgwick try to "verify" the self—evidence of
this principle? Not explicitly; not beyond the claim that

the principle is contained in the notion "ought."
Justice: A Self-Evident Ethical Principle

In Sidgwick's treatment of Justice, it is easier to
see the Intuitive Method at work. It is his treatment of
Justice, too, which first gives rise to a fairly important
question: Is Justice considered a self-evident principle?

Is it argued for? Is it both? If both, does Sidgwick have
a clear sense of self-evidence.

The principle receives several formulations; and these
formulations receive varying kinds of justification. The
fullest discussion, in Book III, Chapter XIII ("Philosophical
Intuitionism") follows the presentation of what I have called
a universalizability principle.

A corresponding proposition may be stated with equal
truth in respect of what ought to be done Epf-not
bye-different individuals. These principles have been
most widely recognized, not in their most abstract and
universal form, but in their special application to
the situation of two (or more) individuals similarly
related to each other . . . in what is popularly known
as the Golden Rule, 'Do to others as you would have
them do to you.‘ This formula is obviously unprecise
. . . for one might wish for another's co-Operation in
sin, and be willing to reciprocate it. Nor is it even
true to say that we ought to do to others only what we
think it right for them to do to us; . . . there may
be differences in the circumstances--and even in the

 

. . . its relation to other notions with which it is connected
in ordinary thought." Sidgwick, Methods, p. 33.

50

natures--of two individuals . . . [Tlhe self-evident

principle strictly stated must take some such nega-

tive form as this; 'it cannot be right for A to treat

B in a manner in which it would be wrong for B to

treat A, merely on the ground that they are two dif-

ferent individuals and without there being any dif-

ference between the natures or circumstances of the

two which can be stated is a reasonable ground for

difference of treatment.
Sidgwick does not make explicit the kind of proof he is using
here; in terms of the account given above it might be claimed
that he is using the "inconceivability" criterion--for he
ends up stating the principle in negative form-~in other words,
that he shows the negation of the principle (as first stated)
to be inconceivable. In the same vein, it might be claimed
that when he considers "Golden Rule" formulations he is also
using the inconceivability criterion, here finding the nega-
tion quite conceivable and thus rejecting the "popular" Golden
Rule formulation.

On closer examination, however, there are some prob-
lems with this interpretation of his procedure. Which sense
of "inconceivable" would be apprOpriate? One sense discussed
was "incredible": i.e., "I cannot believe that it is true."
But there are vast differences between "I cannot believe that
I am not thirsty" and "I cannot believe that what is right

for me to do could be wrong for someone else, in the same

circumstances, to do." Beliefs about one's immediate experience

 

lSidgwick, Methods, pp. 379-80.

51

belong to an epistemological category which William Alston
calls "self—warranted beliefs."l These are beliefs which are
immediately justified; i.e., their justification consists
simply in being held, and no other knowledge is necessary for
their justification. According to Alston-~and I agree-~any
proposition about one's immediate experience which one
believes to be true is justified. Beliefs about morality

are not in the same category; they are not justified simply
by being held. This sense of "inconceivable," then, is not
the one being used.

Another interpretation of "inconceivable" was "I do
not know what it would be like for this to be true." Perhaps
this interpretation is more helpful, although again Sidgwick's
particular example is from a realm far removed from morality.
"I do not know what it would be like for straight lines to
enclose a space." "I do not know what it would be like for
an action to be right for one person and wrong for another in
the same circumstances." This, I think, is the most accurate
interpretation of Sidgwick's procedure in "assuring himself
of the self-evidence" in this particular discussion of Justice.

It is not, however, Sidgwick's only discussion of that
principle. He immediately proceeds to draw parallels between

Justice and another self-evident principle, Prudence.

 

lWilliam Alston, "Self-Warrant: A Neglected Form of
Privileged Access," American Philosgphical Quarterly 13 (Octo-
ber, l976):257-272.

 

52
Justice and Prudence

The principle of Justice now receives this formulation:
"individuals in similar conditions should be treated similarly."
This principle "emerge[s] in the consideration of the similar
parts of a Mathematical or Quantitative Whole."l The principle
of Prudence is another which emerges from mathematical con-
sideration: "a smaller present good is not to be preferred
to a greater future good."2 This mathematical analogy is my
major concern, for it is a means of "assuring oneself of the
self-evidence" of these two principles.

Before pursuing the analogy, however, another problem
must be dealt with. Sidgwick gives several formulations of
the principle of Prudence, and they are not equivalent formu-

lations. Consider:

The proposition 'that one ought to aim at one's own
good' is sometimes given as the maxim of Rational
Self-love or Prudence: but as so stated it does not
clearly avoid tautology; since we may define 'good'
as 'what one ought to aim at.’ If, however, we say
'one's good on the whole,' the addition suggests a
principle, . . . not tautological . . . that 'of

impartial concern for all parts of our conscious
life.’

Granted, Sidgwick does not claim that this last formulation

is equivalent to the first. In the passage cited, the first

formulation "suggests" the second. In the Index of Methods,

 

1Sidgwick, Methods, p. 381.

2Ibid.

3Ihid.

53

under "Prudence," we read that "impartiality" is "educible
from the rule of" Prudence. This entry refers us to the same
page-~381--as the lengthy passage quoted above.1 (Note that
here again there is the question of whether Sidgwick identifies
the self-evident with the underived. What difference can there
be between "educed" and "derived;?)

At any rate, Sidgwick is working with two distinct
principles and for purposes of clarity these should be sepa-

rately named. I will use "P " to refer to the principle, "One

1
ought to seek one's own good." Sidgwick refers to this prin-
ciple as Prudence/or Rational self-love. I will use "P2" to
refer to the principle, "One ought to have impartial concern
for all parts of one's conscious life." Sidgwick often refers
to this second principle as one of "impartiality." Sidgwick's
exposition would be clearer if he explicitly separated, and
officially named, these two facets of the principle of Prudence.
To return to the mathematical analogy and the parallel

between Prudence (P2) and Justice, Sidgwick writes that just
as the notion of

. . . the 'Good on the Whole' of a single individual

. . . is constructed by comparison and integration of

the different 'goods' that succeed one another in the

series of our conscious states, so we have formed the

notion of Universal Good by comparison and integration

of the goods of all individual human--or sentient--

existences. And here again, just as in the former

case, by considering the relation of the integrant
parts to the whole and to each other, I obtain the

 

1Ibid., p. 525.

54

self-evident principle that the good of any one

individual is of no more importance, from the point

of view . . . of the Universe, than the good of any

other . . . And it is evident to me that as a

rational being I am bound to aim at good generally

. . . not merely at a particular part of it.
Sidgwick notes that this principle of Justice does not only
apply to cases where "good" is interpreted hedonistically

it is equally applicable to any other interpreta-

tion . . . in which good is conceived as a mathe—

matical whole, of which the integrant parts are

realiged in different parts or moments of a life-

time.
The question which arose concerning the principle of univer-
salizability also arises here: given these mathematical analo-
gies, is there any clear sense of "underived" built into his
concepts of "self-evident" and of "intuition"? This notion
of the relation of parts to whole is somewhat mysterious.
Although this sounds like an attempt to establish a principle
which is then labeled self-evident (note the phrase "I obtain
the self-evident principle"), it does not accomplish that goal
for the reader. To the extent that there is an argument here,
it runs like this:

1. There are no grounds for preferring one part of a

mathematical whole over other, equal parts.

2. The good of each individual human being is an

equal part of a whole ("Universal Good").

 

1Ibid., p. 382.

2Ibid., p. 381.

55

3. Therefore, the good of one human being is not
preferable to that of any other.

The mathematical first premise offers a very clear
example, even a model, of what "equal parts" means; but the
argument stands without that premise. The argument is in
fact almost tautologous: the good of each human being is
equal to that of every other human being (premise 2); there-
fore the good of one human being is not preferable to that of
any other (conclusion). If "equal" means "equal in value"
and "preferable" means "of greater value"--and these are
plausible interpretations--then the "argument" is circular.

(I am discussing only validity here, not soundness.) The
mathematical parallel is useful as explanation rather than as
part of the proof. Sidgwick's comment that the principle of
Justice emerges from the "consideration of the similar parts
of a Mathematical or Quantitative Whole" remains obscure. It
appears that what real progress has been made in the argument
is from a claim about equal value to a demand for equal treat-
ment--and the connection between the two cannot be established
by mathematical analogies.

Two other, fairly minor, problems surface in these
passages. One is Sidgwick's persistent use of the notion of
"whole"; in the case of Justice, that whole is "Universal
Good," in the case of Prudence, it is the good of one person's
lifetime which is considered as a whole. Is it necessary to

his ethics or to his epistemology that these things be viewed

56

as additive wholes, made up of the sum of their parts?
Couldn't he more simply have said that goods may be compared
with one another and ranked, and that neither temporal posi-
tion nor "ownership"--i.e., belonging to this person rather
than that-~make one good higher than another?

And, as in the case of the mathematical analogies,
there is the question of whether or not Sidgwick's conclusion
(in this case, that future good is, ceteris paribus, equal to
present good) really draws any support at all from Sidgwick's
comments. The principle "aim at one's own good" suggests
"aim impartially at one's own good, regardless of its position
in time" only if it is already held that position in time is
irrelevant to the ranking of goods.

The major point of this discussion, however, is not
the success or failure of Sidgwick's arguments. It has been,
instead, an examination of what Sidgwick means when he calls
a proposition "self-evident." Can his claim that these propo-
sitions are self-evident be squared with his various attempts
to establish their truth? That question is addressed directly
in Section V, drawing upon the explication just completed.

Nevertheless, the question of the success or failure
of his arguments is important. Chapter V looks at what Sidg-
wick has to say about Prudence (P2) and about Universal Good
from a perspective offered by some contemporary ethicists,

and finds Sidgwick's position sound.

57
Rational Egoism and Rational Benevolence

Finally, let us look at the places where Sidgwick
presents the self-evident principles of Rational Egoism and
Rational Benevolence. At the beginning of Book IV, Chapter
II, he spends some time on a comparison between Egoistic
Hedonism and Universalistic Hedonism, claiming that if the
second must be proved, so must the first; because there are

"1 He

times when each conflicts "with present inclination.
drops the question, however, because Common Sense only demands
a proof of the second. It is quite clear in this passage that
Sidgwick does not see the offering of proof as inconsistent
with self-evidence. Notice, for instance, the following com-
ment:

the fact that certain rules are commonly received

as binding, though it does not establish their self-

evidence, renders it generally unnecessary to pggye

their authgrity to the Common Sense that receives
them . . .4

 

When he returns to the question in the final chapter of Methods
it is again to remark that Common Sense supports Rational
Egoism:

It would be contrary to Common Sense to deny that the
distinction between any one individual and any other
is real and fundamental, and that consequently "I" am
concerned with the quality of my existence as an indi-
vidual in a sense, fundamentally important, in which I
am not concerned with the quality of the existence of
other individuals; and this being so, I do not see hgy_
it can be proved that this distinction is not to be

 

1Ibid., p. 418.

2Ibid., p. 419. Emphasis added.

58

taken as fundamental in determining the ultimate end
of rational action for an individual.

In another passage, Sidgwick puts it this way:

If the Egoist strictly confines himself to stating his
conviction that he ought to take his own happiness or
pleasure as his ultimate end, there seems no opening
for any line of reasoning [emphasis added] to lead him
to UniversalistIC"Hedonism as a first principle; it
cannot be proved that the difference between his own
happiness and another's happiness is not for him all-
important.

 

These are the longest treatments of the truth of the principle
of Egoism in Methods of Ethics. The underlined passage in
each shows Sidgwick's willingness to offer support--should it
be necessary and available--for principles he calls self-
evident. In addition, the last two actually offer such support
for the principle of Rational Egoism.

What kind of support is it? The key notion in each
of the two discussions, that each of us "is concerned" in a
special way with his/her own existence, is somewhat ambiguous.
Sidgwick surely means more than that each of us is specially
equipped to know what is good for himself or herself, because
he takes such considerations routinely into account in his
presentation of utilitarianism, and he is presenting Egoism
as a competitor with Utilitarianism. Does he mean, then,
that as a matter of psychological fact each of us is more

deeply concerned about his/her own welfare than anyone else's?

 

1Ibid., p. 498. Emphasis added.

2Ibid., p. 420.

59

He has repeatedly rejected psychological hedonism-~the
doctrine that in all of our actions our primary motive is our
own greatest pleasure--and most of his arguments are also
applicable against psychological egoism in general. It would
nevertheless be consistent for him to hold that what has the
deepest emotional effect on us is our own welfare, distinguish-
ing effect on emotions from effect on our actions; and such a
belief is of course quite plausible. The point of interest,
however, is that Sidgwick draws a connection between that
fundamental concern and the ultimate end of rational action.
It would seem to be a fair paraphrase of his consideration of
Egoism, then, to say: Each of us cares more deeply about his/
her own happiness than about that of others; it is therefore
reasonable that each of us should seek our own happiness above
that of others. As I've paraphrased it, this treatment
involves an inference; furthermore, that inference depends
upon a certain concept of "reasonable end."

To see more clearly what he means here it is useful
to compare Rational Egoism with Rational Benevolence, the
principle with which Sidgwick contrasts it.

In the competing principles of Rational Egoism and
Rational Benevolence, what is at stake is the prOper end of
rational action. Sidgwick holds that both our own self-
interest and the common good are such rational ends. As we
have seen, the grounds for this claim in the case of the

egoism principle seem to be certain psychological facts. In

60

the case of the Benevolence principle, the grounds are
considerably different. Sidgwick approaches this principle
in two quite similar ways. The first approach follows imme-
diately upon the discussion of the principles of Prudence and
Justice. Benevolence is arrived at in an analogous manner.

Here again, just as in the former case, by considering
the relation of the integrant parts to the whole and
to each other, I obtain the self-evident_principle
that the good of any one individual is of no more
importance from the point of view of the universe than
the good of any other . . . and that I am bound to aim
at good generally . . . not merely at a particular
part of it. Fran these two rational intuitions _w_e_
may deduce as a necessary inference, the maxim of
Benevolence . . . that each one is morally bound to
regard the good of any other individual as much as

his own.

 

 

This reasoning is similar to what Sidgwick adduces in support
of the principles of Justice and Prudence. The concept of
"rational end" and the very interesting phrase, "from the
point of view of the universe," remain unclear; these are
dealt with in Chapter V. The interesting point here is that
Sidgwick quite explicitly labels the reasoning process an
inference. And yet there is no doubt that he claims the
principle to be self-evident and intuited; the claim is some-
what modest on this page ("at least a self-evident element,

immediately cognisable by abstract intuition") but later, he

 

1Ibid., p. 382. Emphasis added.

61

refers to Rational Benevolence "as one of the few Intuitions
which stand the test of rigorous criticism."1
The second approach to Rational Benevolence also
involves a reasoning process of sorts. This discussion is
aimed at the Rational Egoist. Sidgwick distinguishes two
strategies on the part of Egoists. The first claims that one's

own good is of supreme importance to oneself; the second claims

that his happiness or pleasure is Good, not only for2
him but from the point of view of the universe . . .

An egoist who uses this second strategy can be shown that "hip
happiness cannot be a more important part of Good, taken uni-
versally, than the equal happiness of any other person." This
is much less clearly an inference than Sidgwick's first approach
was. Perhaps he is using here the procedure which he describes
in "The Establishment of Ethical First Principles." This was
the scrutiny of a supposedly self-evident principle for any
unreasonable limitations in its formulation, the removal of
those limitations, and the subsequent formulation of a wider
principle. In the case of egoism, the limitation which is

seen to be arbitrary is the limitation to my own individual
good. The wider principle is that happiness generally (not

just my own happiness) is to be sought.

 

1Ibid., p. 421.

2Ibid., p. 420.

62

Again the evaluation of this argument will be
postponed (until Chapter V). The point now, as throughout
this chapter, is a methodological one. If we refer again to
Criteria of Truth and Error where Sidgwick characterizes as
"self-evident" those beliefs that in ordinary thought are
certain [without qualificatioh] . . . [and whose] certitude
may be distinguished as primary or independent . . ."l we may
well be puzzled. None of the ethical principles which he
describes as self-evident fits that description. The descrip-
tion would remain valid only if some interpretation could be
given to "ordinary thought" and "primary or independent certi-
tude" which was consonant with the treatment which he gave
those ethical principles. Such an interpretation will be
considered at the end of this chapter, but it will be neces-
sary to consider two other points first.

V. Some Problems in Sidgwick's Concept
of Self-Evidence

 

Several questions have emerged in the preceding pages.
The most important of these are: 1) does Sidgwick clearly
distinguish between the self-evident and the inferred; and
2) just what epistemic value does he grant to such appeals?
In answering the second question, I would like to deal also
with Robert Schultz's charge that Sidgwick confuses epistemo-

logical with psychological certainty.

 

1Sidgwick, "Criteria," p. 437.

63

It is most convenient to begin with the question of
whether or not Sidgwick clearly distinguishes between the
self-evident and the inferred. As noted above, there are a
number of places where he seems to fail to do so; with one
exception, however, I will argue, the procedures which look
like attempts at demonstration are actually just attempts at
explication. Sidgwick himself mentions the distinction which
Aristotle draws between the logically prior and the epistemo-
logically prior; Sidgwick's position might be called a recog-
nition of the difference between the logically self-evident
and the psychologically self-evident. He might remind us that
no statement at all is self-evident as it stands. It is
self-evident only in relation to a reader. Certain conditions
must be met before a statement's "self-evidence" becomes obvi-
ous to a given reader. The reader must, for instance, know
the language in which the statement is written. And then,
the statement must be presented; if someone has never thought
of a certain statement, the question of self-evidence never
comes up. Much of what Sidgwick does in the passages cited
above simply brings to the attention of the reader certain
propositions which he or she may never before have considered.
It is unquestionably true, however, that at times Sidgwick
seems to be doing more than that; when, for instance, he uses
reasoning of this sort: if X is self-evident, then so is y.
Note that he is saying more than "If X is true, then so is y."

It is this kind of procedure which Robert Schultz attacks in

64

his article, "Sidgwick on Proof in Ethics."1 Schultz points
out that the nature of the self-evidence is not made explicit,
and that Sidgwick does not defend his claim that the wider
principle is less arbitrary, more founded in reason, than the
narrower principle. Since some kind of "demonstration" (not
necessarily logical demonstration) is presented, and since
Sidgwick attends to the psychological conditions for assenting
to self-evident propositions, Schultz claims that Sidgwick
confuses normative and psychological accounts of proof. He
refers to two crucially important passages: first, the sec-
tion of Methods which Sidgwick explicitly labels "The Proof
of Utilitarianism"; secondly, the article referred to earlier,
"The Establishment of Ethical First Principles." The first
passage relates more directly to the comparative epistemic
status of "intuitions" and "Common Sense"; I will deal with
it later. The H129 article, however, is directly relevant to
the question of Sidgwick's treatment of self-evidence. First
principles, he writes,

cannot stand in need of what is strictly to be called

proof . . . Yet . . . where there is a conflict . . .

as rational beings conversing with other rationals it

seems absurd that we should not be able to explain to

each other why we accept one first principle rather

than another. And how can these principles be valid

if they do not prove the first principle which they
(to use Mill's phrase) "determine the mind" to accept?

 

1Robert Schultz, "Sidgwick on Proof in Ethics,"
Philosophical Archives, September, 1975, pp. 336-351.

 

2Sidgwick, "Establishment," p. 106.

65

It is, of course, that reference to Mill's phrase which lends
credence to Schultz's claim that Sidgwick is confusing "empiri-
cal" with "normative" concepts of proof. Now, underlying
Schultz's criticism is the assumption that Mill's phrase--and
Sidgwick's reference to it--concerns a purely "empirical" con-
cept of proof. As Schultz puts it, such "proof" would be

simply a general description of the thought patterns

usually found to be present in people's consideration

of moral issues . . . [comparable to] a description,

say, of the thought patterns typically present where

people have arrived at the view that members of other

races are mentally inferior to members of their own

races. In neither case would there need to be any

implications about the truth1 authoritativeness, etc.

of the principle arrived at.
In the first place, this interpretation of Mill's phrase is
questionable. My purpose here is not an exegesis of Mill's
position; instead I would like to point out one plausible
interpretation which would defend Sidgwick against Schultz's
accusation. It seems to me that what both Mill and Sidgwick
are pointing out is this: nowhere in ethics can we find truths
' which are logically certain.2 Sidgwick (and probably Mill,
too) is seeking a kind of certainty which is not logical--and

yet which is still certainty. Whether or not they succeed in

this search is worthy of a great deal of debate; but it is

 

lSchultz, "Proof," p. 346.

2They might have added that neither are such state-
ments found in empirical science; but whether or not the
"certainty" found in the two fields is comparable is a vast
and difficult question.

66

unfair to decide that the only alternative to this logical
certainty in ethics is empirical description of what people

do in fact believe. It is that decision--or assumption--
which underlies Schultz's criticism. It could in fact be
argued that the phrase "empirical proof" is self-contradictory;
any usual use of the word "proof" involves reference to a
norm--in Schultz's words, to the truth, and authoritativeness
of what has been proven.

In the second place, Schultz's description of the
examination of Common Sense Morality drastically oversimplifies
the role that examination plays in Sidgwick's work. Nowhere
is Sidgwick simply taking a survey and listing the results.

It is obvious from even a summary of the three methods of
verification that commonly held opinions are subject to Intui-
tive and Discursive verification. The role which Common Sense
Morality plays in Sidgwick's theory is discussed further in
Chapters III and IV.

But this discussion of Schultz pleaves untouched the
earlier difficulty; does Sidgwick clearly distinguish between
the self-evident and the derived?

In general, I think, any claim that he does not,
depends on the assumption that what is self-evident must be
self-evident to any reader on the first reading; that there
can be no kind of helpful considerations which could result
in a reader's realization that a proposition is self-evident;

that any attempt to explain a principle, especially by relating

67

it to other principles, must necessarily be an attempt to
prove or demonstrate that principle. If my reflections above
are correct, then such attempts are possible, often necessary.
Can such attempts at explication be distinguished
from demonstration? In the latter case the statement under
consideration cannot be known to be true unless premises from
which it is derivable are known to be true. In what I have

called explication, the statement may not be discovered

 

unless the "premises" are known; but its truth can be seen,
recognized, or understood, independently without or before
that of the first. There is no epistemological dependence.
Self-evidence remains a term whose extension is not completely
delineated; it remains a term which is relative to particular
minds, and thus to some extent a psychological predicate.

Let us look at some instances where it might be
argued that Sidgwick was deriving statements and then calling
them self-evident. To begin with, what is the relationship
between the objectivity of ethical judgments and the princi-
ple of universalizability? It is true that Sidgwick points
out a connection between these, and that when he arrives at
the principle, his comments contain the word "therefore."

But it is also true that his claim about universalizability
is that it is implicit in all ethical judgments; what he has
done is to make it explicit. In addition, the connection

which he draws between objectivity and the principle of uni-

versalizability is not a logical connection. It might almost

68

be called a factual connection; at least, it is a description
of what we can and do take into account as rational moral
agents.1 Therefore Sidgwick was neither attempting nor in
fact giving any logical demonstration of the principle of
universalizability. As for the principles of Justice and
Prudence, it has already been maintained above that Sidg-
wick's mathematical analogies are not in fact logical demon-
strations. The question remains, however, as to whether
Sidgwick intended them to be logical demonstrations. A close
look at his text indicates that he did not. The suspicious
phrase is this: "I obtain the self-evident principle." The
"obtain" here is not meant to indicate the final step in a
chain of logical proofs. Instead, it again indicates that a
certain amount of reflection enters into any recognition that
a principle is self-evident. The mathematical analogies again
serve as illustrations of this reflective process; they are
not offered as demonstrations.

What kind of "proof" does Sidgwick offer for Rational
Egoism? None; although the argument as I reconstruct it con-
tains an inference, it is clear in the passages quoted above
that he considers not Rational Egoism, but its denial, as
needing proof. The only comment which might count as the

proffering of evidence is his invocation of Common Sense.

 

lTo Schultz, this and similar passages betray a
confusion between normative and empirical concepts of proof.

69

The relationship between Intuition and Common Sense in
Sidgwick's thought is complex and will be examined in the
next chapter.

Sidgwick's treatment of Rational Benevolence is far
more problematic. He says quite unequivocally that Rational
Benevolence is a necessary inference from the principles of
Prudence and Justice, but he also maintains that it contains
"at least a self-evident element, immediately cognizable by
abstract intuition."

Is it possible to maintain that a principle is an
inference and still maintain that it is self-evident? On at
least two interpretations of inference, yes, this is possible.
First, it is always possible that "inference" is being used
(perhaps improperly) to represent the kind of explication
described above. Secondly (and more probably, in this case)
it may be that the principle in question is a necessary
inference from other principles, but that it could be kpgyg
to be true without them. Take, for instance, the following

mathematical example:

(a) 3 + l = 4
(b) l + l = 2
(c) 4 + 2 = 6
(d)(3+l)+(l+l)=6

(d) is a necessary inference from (a), (b), and (c); on the
other hand it is also self-evident. Rational Benevolence may
be a necessary inference from other intuitions, but it is

also, Sidgwick claims, evident in itself. These other

70

intuitions serve as analogies and explications in the same
sense as the mathematical analogy. For anyone to whom

Rational Benevolence is not self-evident, Sidgwick can call
their attention to the simpler principles without, however,
relinquishing the claim that Rational Benevolence is self-

evident.

VI. Conclusion

 

This chapter begins an exploration of what Sidgwick
means by "intuition." It examines this notion as it occurs
in his Methods of Verification, and concludes that the funda-
mental notion involved in Sidgwick's "intuition" is "self-
evident." Since this self-evidence is something of which we
may "assure" ourselves, the questions arise: What kind of
assurance is available? and, Is this "assurance" different
from "demonstration"? These questions are considered in the
light of those ethical principles which Sidgwick actually
calls "self-evident," and in the light of Schultz's recent
criticism of Sidgwick. The chapter concludes by arguing that
Sidgwick's notion of self-evidence is coherent: that is, his
verifications and explications of ethical propositions are,
in principle and in fact, compatible with his claim that
these propositions are self-evident. The chapter also argues
that his concept of self-evidence is more than a reference to

the psychological persuasiveness of the principles involved.

71

However, it has become clear throughout the chapter
that Sidgwick's distinction between what is intuited and what
is sanctioned by Common Sense needs attention. It has also
become clear that his explications of the several fundamental,
"self-evident," "intuited" notions depends upon an under-
standing of moral reasoning which Sidgwick does not fully

explicate. The next chapters address these two problems.

CHAPTER III

COMMON SENSE

I. Introduction

 

The previous chapter examined Sidgwick's intuitionism,
the intuitionism to which he is forced by his concept of an
adequate moral theory. The analysis showed that, for Sidg-
wick, intuition is the recognition that certain non-
tautologous propositions are self-evidently true. Sidgwick
discusses three Methods of Verification, whose function is to
test the truth of such prOpositions. The Third Method of
Verification is what Sidgwick calls "The Oecumenical Test."
The purpose of the present chapter is to examine his use of
the Third Method and consider whether his concept of "self-
evident” is separable from his concept of "generally agreed
upon."

A close look at Sidgwick's notion of general agreement
reveals some apparent problems. It is not always clear whose
consent is to count; nor what degree of certainty is supposed
to result from general consensus. Sometimes "Common Sense"
seems to refer to an already existing body of beliefs, and
other times it seems to be a method of establishing such

beliefs. Propositions of quite different kinds are subjected

72

73

to the Method of Verification, with different results. The
second section of this chapter discusses these apparent prob-
lems in some detail.

Since the term "Common Sense PhilOSOphy" originates
with Thomas Reid, the third section of this chapter examines
some of his writings, in an effort either to find the source
of Sidgwick's apparent inconsistency, or to find a remedy for
it. The examination shows that the problems identified above
are implicit in Reid's work, and become explicit in Sidgwick's
because the later writer attempts far greater precision.

The fourth and fifth sections of the chapter classify
propositions according to their epistemological level (i.e.,
according to their relative degree of generality and funda-
mentality). The classification helps solve some of the prob-
lems presented in the second section, and allows the unsolved
ones to be formulated more clearly. These unsolved problems
become the subject matter of Chapter IV.

II. Problems Within Sidgwick's Use
of 1'Common Sensefl

 

 

An attempt to pull together the different things which
Sidgwick writes concerning "The Philosophy of Common Sense,"
the "Oecumenical Test," and "The Morality of Common Sense"
reveals a number of apparent inconsistencies and incompletely
worked-out relationships. Some of these problems, commenta-

tors have traditionally explained away (following Sidgwick's

74

lead). Others have been taken to reveal systematic problems
within Sidgwick's epistemology. There are at least four such
problems. One concerns the epistemic status of Common Sense;
a second, the question of whose consent is to count; a third,
the relationship between general assent as a Method of Verifi-
cation and Common Sense Morality; and the fourth concerns the
relationship between Sidgwick's use of these terms and his
understanding of their historical basis in Scottish Common

Sense Philosophy.
Epistemic Status

In the article referred to earlier, Robert Schultz
argues that there is a "fundamental unresolved tension" in
Sidgwick's treatment of the rules of Common Sense Morality.
He points out that a major part of Sidgwick's proof of utili-
tarianism is his claim that utilitarianism "(1) sustains,

(2) harmonizes, and (3) completes the morality of common
sense."1 Sidgwick must, then, grant some epistemic authority
to the morality of common sense; he has to allow some claim
to truth in these propositions, since they are essential pre-
mises in his argmment. This, Schultz claims, results in an
insoluble problem.

The obvious difficulty with this has to do with the

dependence of this whole set of arguments upon the

prior acceptance of common sense morality as authori-
tative . . . [T]he question remains . . . how a

 

lSchultz, "Proof," p. 350.

75

principle that is, in any sense, "argued for" or
"proved from" premises that are the rules of common
sense morality, can be other than a lowest common
denominator of those rules, or can, by itself, be
used to correct and extend these rules.

To some extent Schultz's objection deals with the
coherence of Sidgwick's concept of proof (and, in fact, the
title of his article is "Sidgwick on Proof in Ethics"). This
objection will be dealt with in Chapter V. However, his
claim that Sidgwick is inconsistent concerning the epistemic
status of common sense is directly relevant to this chapter,
and can be borne out by reference to Sidgwick himself. In
"The Criteria of Truth and Error" Sidgwick gives, as his third
method of verification, Oecumenical Verification. Also called
Social Verification, it consists of consensus among those who
are competent to judge. "Competence" is left quite vague.

He here clearly grants epistemic value to consensus; and, to
forestall the objection that perhaps we should distinguish
Social Verification from Common Sense, let me note that Sidg-
wick explicitly identifies the two: "In provisionally taking
Common Sense as the point of departure for philosophical con-
struction, it was this criterion (Social Verification) that
we implicitly applied."2
The word "provisionally" is important here; the third

Method of Verification allows only "provisional" certainty,

 

1Ibid.

2Sidgwick, "Criteria," p. 464.

76

and this is perhaps Sidgwick's recognition of the tension to
which Schultz points. On the other hand, it must be remembered
that Sidgwick is insistent that his Methods merely reduce the
risk of error; he does not claim that any one of them guaran-
tees certainty, and he does not distinguish the provisional
assurance of the Third Method from that of the others. He
intended, in speaking of the tOpic shortly before his death,
to work out such distinctions, for he writes,

At a later period I shall try to coordinate and

compare the different kinds and degrees of imperfect

certitude or provisional acceptance in cases where

this triple verification (i.e., sa isfaction of the

three Methods) cannot be obtained.
In particular he intended to discuss the situations when two
of the three were satisfied, and one was not; ranking the cer-
tainty attained in each combination of two would be implica-
tion, have ranked all three in relation to one another. This
intention, however, was never fulfilled.

So Sidgwick clearly grants epistemic value to general

assent, to Common Sense, although he is careful to call this

value provisional. When Sidgwick examines Common Sense Moral-

ity in Methods of Ethics, he finds that its provisions do not

 

pass the other tests of self-evidence. He considers the

results of his examination almost wholly negative;2 and yet,

 

1Ibid., p. 465.

2Sidgwick, Methods, p. 343.

77

Schultz and others argue, Common Sense Morality contributes

the premises for Sidgwick's major argument in Methods.
Whose Consent Counts?

The nature of this "provisional" certainty, then, is
the first difficulty in Sidgwick's use of Common Sense. The
second is a problem which Sidgwick tries to eXplain away in
terms of the "negative results" referred to above.

It has been fairly urged that I leave the determina-

tions of Common Sense very loose and indefinite: and

if I were endeavouring to bring out a more positive

result from this examination, I ought certainly to

have discussed further how we are to ascertain the

'experts' on whose 'consensus' we are to rely in this

or any other subject. But my scientific conclusions

are to so great an extent negative, that I thought it

hardly necessary to enter upon this discussion.
Sidgwick's commentators have generally accepted this explana-
tion. Schneewind simply paraphrases Sidgwick on the subject.2
(Schneewind'sbook is intended to be as full and consistent an
interpretation of Sidgwick as possible; it is not intended to
be a critical study.) Koch points out a further "negative
result" which makes the specification of whose consent is to

count still less necessary: "the moral reasoning of every

man who attempts to be rational is divided against itself

 

1Ibid., p. 343 n.

2Schneewind, Sidgwick, p. 267.

78

because it allows the reasonableness of both Egoism and
Benevolence."l

Nevertheless, if general assent is to have any episte-
mic value at all--and for Sidgwick, no matter how unclear that
value, it exists--there must be some specification of whose
agreement is to count.

Let us look first, then, at Sidgwick's use of the term
"Common Sense" in Methods of Ethics. Hardly a page of that
book lacks reference to "Common Sense" or to consensus. The
list of phrases which expand on those terms, explain them, or
replace them, is long. The list is also revealing, for it
shows what a variety of ideas is included in that simple
phrase. Such phrases as, "in the view of most men," "ordinary
rules of morality," "natural to most men," "common use," occur
frequently.

Once he writes that it is "the Common Sense of civil-

"2 although it is

ized Europe, upon which we are now reflecting
only a page.later that he reverts to the phrase "Common Sense
of mankind."3 Such explicit ethnocentrism is rare: it is
certainly no more frequent than Sidgwick's occasional explicit

disavowals of ethnocentrism. In his treatment of Benevolence,

 

lDonald Koch, "Henry Sidgwick's Moral Skepticism,"
Manuscript, p. 23 n.
2Sidgwick, Methods, p. 296.

3Ibid., p. 297.

79

for instance, he refers to the customs of other ages and other
cultures and urges his readers to distinguish "real intuitions
of rightness from the blind sense of obligation arising out of
mere custom."l

This passage is not found in the last edition of
Methods of Ethics; it is replaced by some general considera-
tions about obligations to conform to custom. There is no
reason to think, however, that Sidgwick abandons the distinc-
tion which he makes in this passage. The most likely inter-
pretation of his references toward "Civilized Europe" is that,
while he recognizes that for all practical purposes this is
the only kind of consensus which he can know, he also recog-
nizes the dangers of arguing from such limited data. Ethno-
centrism is a practical danger for Sidgwick rather than a
commitment within his theory.

There is another tension, however, which is more
difficult to resolve. It shows up, to some extent, in his
explicit definition of the "Morality of Common Sense": "a
body of moral truth, warranted to be such by the consensus of
mankind, or at least of that portion of mankind which combines
adequate intellectual enlightenment with a serious concern for
morality."2 Apparently the consent of mankind in general is

not.necessary; that of reflective mankind is sufficient. But

lSidgwick, Methods, p. 218.

2Ibid., p. 215.

80

outside of Methods of Ethics some vacillation appears. In
several places Sidgwick refers to the consensus of experts1
and at least once explicitly discounts the opinions of those
who have not thought seriously about the matter in question.2
Is there some way to reconcile "consensus of experts" with
"Common Sense of mankindf? Perhaps there is, if we are will-
ing to limit the latter phrase to those who think seriously
about the subject--in this case, about ethics. If, then, we
call any person who has reflected on ethical questions an
"expert" on the subject, we will have resolved the tension.
The underlying assumption--and I do not know that Sidgwick
makes the assumption explicit at any time--is that ethical
knowledge is different from, say, scientific or theological
knowledge; different in that it can be reached without special
study by an average, but careful, thinker. The contrast with
theology shows up in a curious, brief article by Sidgwick
entitled "Dialogue on Time and Common Sense." The two charac-
ters are debating on whether God, if he exists, does so outside
of time. One speaker invokes the authority of the ordinary
person.

'The good people who go to church believe in an ever-

lasting deity, enduring through Time, not out of Time.‘
'Yes,' [Sidgwick] replied, 'but I understand that the

 

lSidgwick, "Criteria," p. 464.

2Sidgwick, "Dialogue on Time and Common Sense," in
Lectures on the PhilOSOphy of Kant, ed. James Ward (New York:
MacMiIlan, 1905): P. 402.

 

81

better opinion--as lawyers say--among students of

theology is that the efflux and succession Of Time

takes place only for finite beings . . . I respect

this preponderant Opinion'
The "better Opinion," then, is not that Of the ordinary per-
son. And yet, opinions of such ordinary persons are of some
importance, even in theology. In an interesting analogy,
Sidgwick writes that the "normal mind"--the ordinary person's
thoughts--

move about these questions in a spiral way . . . the

philosopher may save himself unnecessary motion, by

keeping nearer the axis of the spiral . . . he must

try to get the whole of our normal thought free from

confusion and contradiction . . .2
This same approach is recommended for the study of ethics. At
the least, then, in many fields, "normal thought" points the
way; it is Of heuristic value. And, although it doesn't sur-
face in this dialogue, there does seem to be an underlying
assumption that in ethics, unlike most other fields, the
ordinary person who becomes reflective, becomes an expert;
i.e., the consensus of the Opinions of these people counts.

In his discussion of the Scottish philosopher Thomas

Reid, from whom Sidgwick borrowed the phrase "Common Sense,"
he gives still another specification of whose consent is to

count. Sidgwick mentions the dispute between Hume and Reid

and reminds us that Hume has One set of beliefs in the

 

1
p. 401.
2

Sidgwick, "Dialogue on Time and Common Sense,"

Ibid., pp. 402-403.

82

"philosopher's closet" and a different--in fact inconsistent--
set in the ordinary world of

'dinner, backgammon, and social talk.‘ Reid takes
his stand with those who are 'so weak as to imagine
that they ought to have the same belief in solitude
and in company!‘ His essential demand, therefore,
on the philosopher, is not primarily that he should
make his beliefs consistent with those Of the vulgar,
but that he should make them consistent with his

own . . . it is the duty of a philOSOpher--his duty
fig a philosopher--to aim steadily and persistently
at bringing the common human element of his intel-
lectual life into clear consistency with the special
philOSOphic element.

 

There emerges here a new phrase in explication of "Common
Sense": "the common human element Of [one's] intellectual
life." Sidgwick focuses here not on how many people--common
people or experts--hold the Opinion, but on the fact that the
philosopher himself holds it, and should not pretend that he
does not.

Two other possible problems need to be mentioned here.
One arises from Sidgwick's comment, in the same article on
Reid, that the ordinary person may not be capable of formulat-
igg these principles, but is Often capable of judging their
truth; "just as few of us could have found out the axioms
required in the study Of geometry, but we could easily see the

2

truth Of Euclid's." And here Sidgwick distinguishes between

two kinds of proposition: those which are the underlying

 

lSidgwick, "The Philosophy of Common Sense,‘
tures on the Philosophy of Kant, pp. 415-16.

2

in Lec-

 

Ibid., p. 417.

83

assumptions of arguments, and those which are its conclusions.
The assumptions, he writes, are "implicitly assented to by
men in general." It is, of course, still harder to determine
what counts as assent if that assent need only be implicit.

(Such a qualification also occurs frequently in Methods of

 

Ethics.)

The other additional problem in this area of "whose
consent is to count" concerns what kind Of explicit gigsent
is to count. As Koch remarks,

If, for example, the average wage earner is casually
asked if he ought to get a higher wage, he will
undoubtedly reply in the affirmative . . . If . . .
some basis is utilized to limit dissent to 'serious
dissent' or 'dissent that really counts,’ 'then howl
is the adequacy of this basis itself to be judged?‘

The Relationship Between the Third Method of
VerificatIOn and Common Sense Morality

 

A number Of problems, then, arise from attempts to
clarify "common" sense or "general" assent. Still another
kind of problem is revealed by the apparent interchangeability
of the two phrases just quoted. It will be remembered that in
"Criteria of Truth and Error" Sidgwick links the Third Method
of Verification (the Oecumenical Method) with Common Sense by
calling an appeal to the latter, an implicit application Of
the former. It is nevertheless true that in the section of

Methods devoted to the examination Of Common Sense Morality,

 

1Koch, "Sidgwick," pp. 26-27.

84

general assent is one of the methods used to Egg; the
Morality of Common Sense for self-evidence.l There is a
kind Of dialectic here, for as long as the maxims of Common
Sense

are left in the state Of somewhat vague generali-

ties . . . it may be fairly claimed that the assent

is approximately universal . . . But as soon as we

attempt to give them the definitiveness which science

requires we find that we cannot do this wighout

abandoning the un1versal1ty of acceptance.
It might be claimed that only the "vague generalities" count
as the Morality of Common Sense, and that therefore there is
no difference between that and the Oecumenical Test. For
various reasons, however, I believe that difference is real
and significant. For one thing, Sidgwick does not confine
the term "Common Sense Morality" to that body of vague gener-
alities; for another, general assent _i__§_ used as a test Of
the self-evidence of the Morality Of Common Sense and it is
difficult for something to be a test of itself. The most
significant difference, however, is implied in the previous
point: the Oecumenical Test, or the test of general assent,
is a procedure, while the Morality of Common Sense is a set
Of propositions. It is the relationship between the two--a

relationship which Sidgwick comes close to calling identity,

but which clearly cannot be that--which needs further attention.

 

lSidgwick, Methods, pp. 341-342.
2Ibid., p. 342.

85

The examination of this relationship will help to clarify

Sidgwick's concept of Common Sense.

Methods vs. "The Philosophy of Common Sense"

There is one more important problem in Sidgwick's use
of the term, a problem which emerges from a comparison of

Sidgwick's use of the term in Methods of Ethics, with his

 

discussion of the concept in an article entitled, "The Philo-
sophy of Common Sense." In this article Sidgwick gives an
initial definition of Common Sense as

the body of unreasoned principles of judgment which

we and other men are in the habit oflapplying in

ordinary thought and discourse . . .
He goes on to list, as typical Of those "unreasoned princi-
ples," four beliefs: that Objects Of sense perception exist
independently Of our perception; that there is a permanent
subject of our changing conscious states; that right and
wrong can be known and are distinct from considerations of
self-interest; that every change has a cause. As I will
argue later, each of these propositions is, essentially an
assumption, a premise rather than a conclusion. This charac-
terization fits in well with Sidgwick's claim, elsewhere, that

such principles are not explicitly formulated by most people,

but receive assent when they are formulated.

 

lSidgwick, "The PhilOSOphy of Common Sense," in Lec-
tures on the Philosophy of Kant, p. 415.

 

86

What is the epistemological nature of the propositions
which Sidgwick, in Methods of Ethics, claims are supported by
Common Sense? To what extent are they like the assumptions
for which he allows Reid to claim such support?

In the Preface to the First Edition Sidgwick makes

the simple assumption (which seems to be made

implicitly in all ethical reasoning) that there is

something under any given circumstances which it

is right or reasonable to do, and that this may be

known.
This is a more careful statement of one Of the four "unrea—
soned" assumptions whose authority Sidgwick recognizes in "The
Philosophy of Common Sense." But in Methods he does not char-
acterize this as a belief Of Common Sense, and in fact he
makes no attempt to prove the proposition; it is, as he says,
a simple assumption. Where Sidgwick ggeg use Common Sense as
evidence of truth or plausibility, he uses it to support pro-
positions of a different kind.

His major examination Of Common Sense occurs in Book
III of Methods of Ethics, "Philosophical Intuitionism." Most
chapters of that book examine specific virtues, "Justice,"
for instance, or "The Intellectual Virtues." Each of these
chapters attempts to explicate the Common Sense notion of the
virtue in question; and most of them end with the claim that

the Common Sense notion cannot be defined precisely. Ordinar-

ily (as mentioned above), the more precisely the virtue is

 

lSidgwick, Methods, p. v.

87

defined, the less universal its acceptance will be; the less
precise its definition, the more universal its acceptance.
That is, there is a group of actions which it is uni-
versally agreed are just; another group whose injustice is
universally agreed to; and a third group about which there is
disagreement. The more complete the attempt to classify the
actions in this third group, the less likely we are to find
universal agreement with the classification. Common Sense,

then, is used in Methods of Ethics to verify propositions like

 

the following: "These actions are contrary to the virtue Of
Justice."

How do these propositions compare with those which
Sidgwick permits Reid to support through appeals to Common
Sense? They are not of the same kind as the four which Sidg-
wick explicitly allows. Each of those four is an assumption
which makes possible a whole field Of inquiry. The assumption
that every change has a cause underlies all scientific inquiry,
for scientific inquiry asks what particular-cause brings about
particular changes. The assumption that there is a permanent
identical subject Of changing conscious states underlies,
makes possible, most statements of psychology, for most
psychology asks what is true of that underlying subject. The
assumption that what we perceive exists independently Of our
perception underlies all efforts to distinguish illusion from
reality. Finally, as noted above, the assumption that there

is a knowable difference between what is right and what is

88

wrong underlies more particular inquiries about whether
particular actions or classes of action are in fact right.

But the statement that specific kinds of action are
just is not as general, not as fundamental, as these four
assumptions. (There is this similarity in epistemological
status: such a statement does make possible still other
inquiries, e.g., whether a particular action belongs to the
kind of action which is just.) But it is not as general as
those four assumptions because it describes one kind Of right
action among many. It is not as fundamental as those four
assumptions because it is made possible by one of them--the
fourth--that there is a knowable difference between right and
wrong.

Now Sidgwick gives us no examples of what he would
call the illicit use of common sense; he simply writes that
Reid misuses "vulgar common sense" when he allows the "ridicu-
lousness" of conclusions to count as good reason for disbeliev-
ing them. An interesting question, then, is how Sidgwick
distinguishes the proper use of appeals to consensus from the
improper use. On questions concerning assumptions as funda-
mental as the four listed above, Sidgwick treats the appeals
to consensus as entirely proper. When the propositions are
of other sorts--i.e., assumptions, but less general ones; or
conclusions of arguments--then Sidgwick is much more wary of
use Of appeals to consensus. There is no attempt, either in

"The Philosophy of Common Sense" or elsewhere, to separate

89

these other propositions into classes such that one class
would be properly supported by common sense, and the other
not.

There are, then, a number of apparent inconsistencies
in Sidgwick's use of terms like "Common Sense." Since that
term originates with Reid, and since the most important Of
the problems arises in Sidgwick's discussion of Reid, it is
appropriate here to look more closely at what the Scottish

Common Sense philOSOpher meant by the term.

III. Reid and Common Sense

 

In examining Reid's concept of "Common Sense" the
question of interest is whether the apparent inconsistencies
in Sidgwick's concept arise because of confusions in Reid.

If Reid is clear where Sidgwick is not, then the confusions
might be solved by reinterpreting Sidgwick along Reidian lines.
In that case, it could also be asked why Sidgwick diverged
from Reid, and answers could throw further light on Sidgwick's
position. If, on the other hand, these confusions have their
origin in Reid, then we will understand part of the reason

for Sidgwick's problems, and will know that we must look else-
where for a solution. This comparison will refer to Essay VI
of Reid's most important work, Essays on the Intellectual

Powers of Man.1

 

 

lThomas Reid, Essays on the Intellectual Powers of
Man (Cambridge, Mass.: The M.I.T. Press, 1969).

90

This discussion will begin by pointing out the
important similarities between Reid and Sidgwick, and then
will consider whether the specific problems which have been
pointed out in Sidgwick also occur in Reid.

Fundamental Agreement Between
Sidgwick and Reid

TO begin with, it is clear that Sidgwick and Reid are
in substantial general agreement in their use Of the term,
"Common Sense." Like Sidgwick, Reid is interested in setting
up‘criteria for the identification of the intuitive founda-
tions of knowledge. For Reid, as for Sidgwick, the question
of whether such foundations are innate is not important.1
For Reid, too, it is important that these intuitive founda-
tions be non-tautologous.2 Again, like Sidgwick, Reid believes
that it is possible to reason concerning these self-evident
principles. Reid is in fact more explicit concerning this
point, and he describes four particular weapons which philo-

SOphers use to defend their first principles. The first of

 

lReid writes repeatedly in the Essay that we have
these convictions "by the constitution Of our own nature" (p.
540) so that someone lacking belief in our power of reasoning
“would be a greater monster than a man born without hands or
feet " (p. 630). He also insists, in arguments similar to
Hume's, that these convictions cannot arise from experience
(pp. 654, 666). In spite Of this, he maintains that these
principles are not innate, and concludes, "I maintain only
that when the understanding is ripe, and when we distinctly
apprehend such truths, we immediately assent to them" (p.
683). Reid, Intellectual Powers.

2Ibid., p. 479.

91

these is ridicule, i.e., claiming that the denial of these
principles, while not a contradiction, is an absurdity. Ridi-
cule, or the unsubstantiated claim "That's absurd!" is hardly
a weapon Of reason, and Sidgwick takes Reid to task for his
occasional misuse of this "tool." The second is the use of
particular examples, for these will win assent from the common
man more readily than general principles.1 The third weapon
is called argument "ad absurdum."

We trace the consequences Of that supposition . . .

and if we find any Of its necessary consequences to

be manifestly absurd, we conclude the supposition

from which it follows to be false.
This is not just a reiteration of the "argument" from ridicule;
its force comes from Reid's pointing out that

there is hardly any prOposition . . . that stands

alone and unconnected. It draws many others along

with it in a chain that cannot be broken.3
This leads to Reid's most convincing argument, which he calls
an argument "ad hominem." "If it can be shown that a first
principle which a man rejects, stands upon the same footing
with others which he admits . . . he must be guilty of an
inconsistency."4.

This discussion reveals two more fundamental points

of similarity between the two men. It is clear that Reid,

 

1Ibid., p. 662.

2Ibid., p. 608.

3Ibid.

4Ibid.

92

like Sidgwick, is often appealing to some kind of implicit
consent;:L that is, the canmon person is not capable of formu-
lating some Of these principles, but would agree to their
truth upon hearing the principle formulated by someone else.
The most fundamental point of agreement is this: that such
universal consent, whether or however verified, is evidence
for the truth of the prOposition. This is the heart of Common

Sense Philosophy.
Problem Areas in Reid and Sidgwick

The first of the four problems identified in Sidgwick
concerns the epistemic status of Common Sense (above pp. 74-77).
It is "provisionally" authoritative, but "provisionally" is
left unclear. The problem in this form does not occur in
Reid, who takes the findings of Common Sense to be conclusive;
indeed, it is the bedrock of inquiry. Why does Sidgwick part
company with Reid here? It might be argued that Sidgwick's
position, although more problematic than Reid's, is neverthe-
less an advance over it. The later writer advances a number
of methods of examining apparently self-evident truths, a
reaction in part to his recognition that what is called Common
Sense may be simply prejudice in disguise. Sidgwick in fact
takes Reid to task (as noted earlier) for not recognizing this

danger. A major reason for Sidgwick's reluctance to claim

 

lSchneewind, Sidgwick, p. 65.

 

93

complete certainty for Common Sense is that his close
examination of Common Sense morality reveals its inadequacy.
Reid makes no such examination. Let us say, then, that
although Reid is consistent in what he claims about the
epistemic status of Common Sense, the inconsistency in Sidg-
wick's work arises from problems in Reid's position which the
Scottish philosopher did not himself recognize.

The second problem identified earlier in Sidgwick is
that of specifying whose consent counts. Again, this problem
does not occur as explicitly in Reid; he maintains that a
large number Of propositions receive consent from all those
"with whom we can converse and transact business."1 This
formulation is repeated frequently; Reid sees no problem in
identifying such people, and remarks that "all civilized
peOple distinguish those who have this gift Of Common Sense

2 Thus there is no tension between

from those who have it not."
"experts" and the ordinary person, between the reflective and
the non-reflective. It might, Of course, be claimed that the
borderline between "those who have this gift" and others is

not a precise one; but this is a different problem (for Sidg-

wick seems to vacillate on the criteria to be used rather

than on the precise applicability of the criteria).

 

1Ibid., p. 557.

2Ibid., p. 559.

94

That Reid's claims to such universal consent are
sometimes unjustified is unquestionable. It might be that
ethnocentrism is a much more serious danger for Reid than for
Sidgwick. For one thing he frequently claims that certain
principles are found in "all languages"--but refers to nO
language at all (not even English) to substantiate this claim.
Reid makes an unusual distinction between contingent and
necessary first truths. The details of this distinction will
not concern us here,1 except to note that necessary truths are
in some sense stronger than contingent truths. This is Of
interest because this classification contains, along with
logical and mathematical truths, "grammatical" truths, truths
of taste, and those containing morals.2 These latter cate-
gories (and particularly "truths of taste") indicate that Reid
was unaware Of the dangers of ethnocentrism. Sidgwick, on
the contrary, was aware Of the danger and tried, with varying
success, to avoid it. But neither man maintains a theory
which is, as such, ethnocentric.

Again, then, a problem (in this case, that of whose

consent is to count) occurs in Sidgwick which does not occur

 

1The distinction is drawn along somewhat Leibnizian
lines. At one place, for instance, Reid writes that "con-
tingent truths . . . may be true at one time, and not at
another," while necessary truths are always true. Both cate-
gories Of statement, however, are certainly true.

2The contingent truths refer to the reliability of
human faculties such as memory, sense, and intellect, as well
as our beliefs that the future will be like the past, that we
have free will, and that other minds exist.

95

in Reid; and again, the difference results because Sidgwick's
position is more cautious. If Reid is completely consistent,
it is at the price of credibility; if Sidgwick is less than
completely consistent, it is because of a more plausible
position.

The third problem identified in the initial discussion
of Sidgwick was the relationship between the Method of Verifi-
cation called the Oecumenical Test, and Common Sense Morality.
At some places these two are identified; in other places the
first is used as a test of the second. This is yet another
instance of a problem resulting from a more SOphisticated
position; for it is Reid's position that whatever is part of
Common Sense is true, unquestionably true, and not subject to
further test.

The final problem which the earlier discussion pointed
out in Sidgwick does not follow the pattern Of the first three.
It does occur in Reid's writings, to at least the same extent
that it appears in Sidgwick's. This fourth problem concerns
the varying epistemological levels of the propositions which
are said to be verified by Common Sense. When Sidgwick is
discussing Reid, he considers prOpositions which are funda-
mental to whole fields of inquiry, propositions such as "every
event has a cause." When Sidgwick examines Common Sense
Morality, however, he examines prOpositions such as "this
action violates justice," which are not nearly as fundamental.

Reid groups "Truth of taste" with "Mathematical Truths," and

96

the second are not nearly as general as the first. He allows
wide jurisdiction to Common Sense:

It is possible, that, by setting out from false

premises, or by an error in reasoning, a man may be

led to a conclusion that contradicts the decisions

Of Common Sense. In this case, the conclusion is

within the jurisdiction of common sense, though the

reasoning on which it was grounded be not . . .1
These comments, of course, are pointed most directly at philo-
sophers like Hume, who argue from quite involved premises to
conclusions which violate common sense at the most fundamental
level; such philOSOphers tell us, for instance, that inductive
arguments yield no knowledge at all. Nevertheless Reid's
comment indicates no limitation on what kind of proposition
is subject to the judgment of Common Sense, and reveals that
the distinctions unrecognized by Sidgwick are also unrecognized
by Reid.

This examination of the historical background of Sidg-
wick's use Of the term I'Common Sense" shows that the problems
within Sidgwick's work are all latent within Reid's work;
Sidgwick does not diverge from Reid, but makes his terms more
precise. It is this attempt at precision that gives rise to
difficulties within Sidgwick's work. The next section of this
chapter will present a framework for understanding what Sidg-

wick meant by Common Sense, a framework whose purpose is to

resolve these difficulties and show that Sidgwick's position

 

1Reid, Intellectual Powers, p. 568.

97

can, after all, be made consistent. It also intends to
illuminate the relationship between Sidgwick's use Of Common

Sense and his intuitionism.

IV. A Clarifyinngramework

 

I will argue here that it is helpful to think of Sidg-
wick's concept of Common Sense primarily as a procedure, and
that the unclarities discussed above arise from his failure
to recognize that at times he is using Common Sense to mean a
procedure (the Oecumenical Test) and at times to mean a body
of principles ("the body of unreasoned principles of judgment
which we and other men are in the habit of applying in ordinary
thought and discourse.")1 The following formulation identifies
both components of the concept: The appeal to Common Sense is
a procedure which, on different epistemological levels--
different levels of generality--yields different results and
therefore has different epistemic value. Before distinguishing
these levels, I will give a general description of this pro-
cedure.. At every level this procedure is what Sidgwick calls
the Oecumenical Test: the appeal to general assent, which
Sidgwick considers to be a fundamental Method of Verification.
Like the other methods it is only a negative test of certainty;
where such consensus is lacking, so is certainty. Like the

other tests, it is a heuristic device, for when a proposition

 

1Sidgwick, "Dialogue," p. 415.

98

fails the test, we know that we must keep on looking-~for a
new formulation, perhaps a more limited one, which will pass
the test. The subject matter of this test, on every level,
is a set of principles which peOple assume to be true, to
which people appeal in ordinary conversation, although these
principles have often not been explicitly formulated by the
speakers.

These principles can be distinguished from one another
by their relative epistemological priority; by such a cri-

terion, they fall into three different categories.

The First Level: CS1

The first category, identified above in Sidgwick's
discussion of Reid, contains those prOpositions which are the
most fundamental. They are the presuppositions of whole
fields of inquiry. To use Quine's metaphor, they are more
central to our web of belief than any other non-tautologous
statement(s). The only ethical judgment among those we have
considered which belongs in this category is that there is a
knowable difference between right and wrong. A correlative
judgment--that there is a knowable difference between good
and bad--also belongs here. These statements come closest
to passing the test of general assent, and it is easy to
Specify whose consent is to count: everyone's, including the
philosopher herself. There is no danger of ethnocentrism

here, and no distinction between the expert and the non-expert.

 

99

There may be a distinction between the reflective and the
non-reflective, but only to this extent: that for a few
peOple the assent is explicit, for most implicit. Only when
a prOposition is formulated and understood can explicit con-
sent be given, and in most peOple's lives such fundamental
propo itions are never presented.

The epistemic status of these prOpositions is stronger
than that of any other category. While Sidgwick would still
insist on calling the certainty "provisional," there is no
fompetition here. Their certainty is not subject to test by
propositions Of any other category. They are of course sub-
ject to the other two Methods of Verification; this raises the
interesting question of whether these methods can be phrased
as self-evident principles. If, in the form of principles,
they would pass the test Of general assent, further questions
would arise about ranking within this most fundamental cate-
gory. The second method, rephrased as a principle, is almost
analytically true: True principles must be mutually consis-
tent. The first method, in any of its guises, is not analyt-
ically true. Consider, for instance, "Those prOpositions
whose negations are inconceivable are true." This is not a
tautologous claim, nor does it seem self-evidently true. The
same is true of the Third Method itself, rephrased as a prin-
ciple: "Those propositions which receive universal assent
are true." It seems clear that neither of these is self-

evidently true, and that neither would pass the test of

100

universal assent. Sidgwick's wisdom in calling them Methods
rather than Principles is apparent.

I will call the body of principles which have the
epistemological status described above CSl' Those on the

second level, to be described next, will be called C82.

The Second Level: CS2

This level corresponds to what Schultz calls "abstract
transcedent axioms."l The belief that such axioms are true is
what Sidgwick calls "PhilOSOphical Intuitionism."2 This level
includes the principles of Justice and Prudence, of Rational
Egoism and Rational Benevolence. Although I do not know that
Sidgwick ever refers to these in terms of Common Sense, there
is a two-fold warrant for labeling them C82. First, as noted
in the previous chapter, Sidgwick's justification of these
principles is occasionally difficult to distinguish from an
appeal to general assent (and such an appeal, as noted before,
Sidgwick explicitly links with appeals to Common Sense).
Secondly, the principles fit epistemologically between the
more self-evident axioms and those specific beliefs which, in
Methods, Sidgwick explicitly labels Common Sense.

By "epistemological fit" I mean that the principles

of CS2 are less general and fundamental than those of C51,

 

lSchultz, "Sidgwick," p. 337.

2Sidgwick, Methods, pp. 373 ff.

101

but more general and fundamental than those of CS As I

3.

argue above, each of the propositions of CS makes possible

1
a whole field of inquiry. This is not true of those of C82'
Each of them, however, is definitive of a major class of moral
duties. That "individuals in similar situations should be
treated similarly," for instance, cannot be true unless "there
is a knowable difference between right and wrong" is true.
This logical dependence is not reversible; the second proposi-
tion need not be true for the first to be true. The first
proposition, in other words, applies to a broader range of
actions than the second; it is also presupposed by the second.
These principles (Justice, Benevolence, and so on)
receive universal assent when they are made explicit.‘ They
are made explicit, however, only by philosophical reflection;
and so the question Of whose consent is to count is answered:
"by those who have given morality sufficiently careful thought
that they understand the principle." "Understanding the
principle" may be taken to mean understanding the kinds of
considerations which Sidgwick presents in explanation Of the
axioms; and understanding, to some extent at least, the alter-
natives to these formulations and the reasons for Sidgwick's
acceptance of them. There is the classical temptation, of
course, of claiming that anyone who disagrees just hasn't
understood; I am not imputing that position to Sidgwick. There
is a particular problem at this level, however, of specifying

what is to count as sufficient reflection. Certainly it

102

demands more than simply understanding what the sentence
means; it demands some critical reflection upon the implica-
tions of accepting or denying the sentence. This description
is, Of course, not at all precise: and an additional diffi-
culty arises from the likelihood that more reflection is
required at this level than at either of the other two. On
the one hand, their generality makes the propositions Of CS2
more abstract than those of CS3; they cover a wide variety of
actions which differ from one another in many more ways than
they resemble one another. It is likely that the propositions
of CS2 demand more reflection than the latter. On the other
hand, they are more involved and less immediately evident than

those of CS and thus, again, demand more reflection. In this

1
respect--the amount of reflection required for a proposition
to be understood--the epistemological progression from level
one through level three does not hold. On the other hand,
the differing amount Of reflection required is a clue to the
difference in epistemic status between the first and the
second level. It seems clear that Sidgwick considers the
second level somewhat less certain than the first; it is
propositions of the second level that receive his careful
scrutiny and support, while those of the first he feels com-
fortable in simply assuming. Since he does not claim more
widespread assent for one level than for the other, the fact

that less reflection is required probably explains the greater

certainty attributed to propositions of the first level.

103

The Third Level: CS3

Given this tentative characterization of CS1 and C52,

it is time to turn to CS3. The prOpositions of this level

are those which are labeled "Common Sense Morality" in Methods

of Ethics. The belief that such propositions are true is what
1

 

Sidgwick calls "Dogmatic Intuitionism." (What he calls
"Perceptional Intuitionism" would be a fourth epistemological
level, and is not considered here.)2 Propositions of the
third level are less general and less fundamental than those
of the other two levels. For example, the following proposi-
tion is a piece of Common Sense Morality and hence belongs to
C53.

A. Taking what belongs to another is wrong.

If A is treated as cognitive, as a prOposition which
is capable of being true or false, then anyone who claims that
A is true presupposes the truth of the following, from €51:

B. There is a knowable difference between right and
wrong.

Proposition It cannot be true unless proposition B is
true. This order is not reversible; the truth of B does not

presuppose the truth of A. A is also less general than B,

since it applies to a proper subset of the situations to which

 

1Ibid., p. 101.

21hid.. pp. 100, 102.

104

B applied. If it is granted that these propositions are
typical of their levels, it follows that prOpositions Of the
third level are less general and less fundamental than those
of the first level.

It is somewhat harder to establish this epistemo-
and it is the relative

logical order between CS and CS

3 2’
epistemic status of these two levels which is pointed out as

a problem in the second section of this chapter. To begin
with, however, it is reasonably evident that propositions of
the third level are less general than those of the second.

One reason to believe this is true is that there are far more
third level propositions than second. Although Sidgwick
phrases each of the "abstract transcendent axioms" (second
level prOpositions) in several different waysflit is fair to
assume that Sidgwick intended to give at least three but at
most five. I have identified four: Prudence, Rational Egoism,
Justice, and Rational Benevolence. On the other hand, Book
III of Methods, which examines Common Sense Morality (third
level propositions), identifies a large number of separate
duties: Benevolence, Justice, Obedience to Law, Truth-Telling,
Liberality, other Social Virtues, and Self-Regarding Virtues
such as Prudence, Temperance, and Purity. There is more

involved, however, than simply this difference in numbers.

 

lSchneewind identifies a total of seventeen. Sidgwick,
pp. 295-297.

105

Each Of the second level propositions is a constraint on all
moral action, although sometimes the requirement is met vacu-
ously (i.e., the requirement is met simply because it is not
violated). But the third level prOpositions each refer only
to a subset of moral actions. Notice that Prudence, Justice,
and Benevolence occur in both lists.

As a second level proposition, both Prudence and
Justice refer to the rational end of all action. On the third
level, Prudence refers to the habit of careful calculation in
matters concerning our own welfare; on the third level, Benevo-
lence refers to specific acts Of kindness.1 The second level
axioms, then, are more inclusive than the third level prOposi-
tions.

In considering whether CS is epistemologically more

2

fundamental than CS3, it turns out that the clearest answer is,
again, that at least Sidgwick must think them to be. It is
here that Schultz charges Sidgwick with a "fundamental unre-
solved tension." Briefly, Schultz argues that since Sidgwick
"proves" Utilitarianism by arguing that it harmonizes and
completes Common Sense Morality, he cannot consistently claim
that Utilitarianism is Of superior validity to Common Sense.
In Schultz' words, how can a principle that is ”'proved from'

premises . . . be used to correct and extend those (premises)."l

 

lSidgwick, Methods, pp. 329, 238.

2Schultz, "Proof," p. 350.

106

This part of Sidgwick's work is called by Schneewind the
"Dependence Argument." Since he gives Sidgwick what is prob-
ably the most favorable possible interpretation, I will use
that interpretation here. The essential disagreement can be
put in this way: Schultz claims there is a circular rela-

1 that

tionship in Sidgwick's position, between CS and CS

2 3‘
each is, to some extent, used as a premise in the proof of
the other. Schneewind claims there is a one-way relationship;

that whatever validity CS has is dependent upon the validity

3
of C32’ and that CS2 is established on independent grounds.
Briefly, Schneewind finds in Sidgwick a concept of rationality
which demands that any ethical system have both the form and
the content of the abstract transcendent axioms which make up
the second level. Chapter IV of this dissertation argues that
Schneewind's interpretation of Sidgwick is the correct one;
for my purpose here it is sufficient to note that under one
plausible interpretation Sidgwick finds CS2 epistemologically

more fundamental than CS that is, he holds that the proposi-

3;
tions of the third level cannot be true unless those Of the
second level are true, and he does not hold that this rela-
tionship is reversible. Since it was established earlier that

the propositions Of CS3 are less general than those of C52,

and less general and less fundamental than those of C51’ the

 

1For the sake of brevity, "CS " will mean "the set of
prOpositions Of the second level." "683" will mean "the set
of propositions of the third level."

107

claim of relative epistemological priority among the three
levels is tentatively established.

Correspondingly less certainty is established by con-
sensus at this third level. It is here that the fruits of his
search for consensus are so meager that he can say that ques-
tions of whose consent is to count are irrelevant.

My . . . conclusions are to so great an extent nega-
tive that I thought it hardly necessary to enter upon
this discuSSion.

(That is not to say, however, that these prOpositions
have no epistemic value at all. Schultz is quite right in
saying that Sidgwick attaches some such value. Chapter IV
argues that,while the second level propositions are, possibly,
known intuitively, those Of the third level are not. It does
not follow that third level propositions are not knowledge
claims; rather, their claims to be knowledge have to be vali-
dated through reference tO second level prOpositions.)

At the third level more than at the others, such assent
as exists is as likely to be explicit as to be implicit. Peo-
ple frequently speak of an action or a class of actions as
being wrong; and they just as frequently use such principles
as implicit norms in evaluating actions. But, as noted earlier,
agreement at this level depends upon the imprecision of the

formulation. As the formulation becomes more precise, agree-

ment lessens.

 

lSidgwick, Methods, p. 343 n.

108

It was at this level, too, that the introductory
segments Of this chapter found a problem: what is the rela-
tionship between Common Sense Morality and the third method
of verification? At this point it can be seen that an appeal
to Common Sense Morality is, as Sidgwick said, an implicit
use of the third method; but, (and this he didn't recognize)
it is a use of that method at a particular epistemological
level. At that level comparatively little consensus can be
reached. The examination of such prOpositions by the third
method shows, in detail, to what extent such agreement does

exist.
Conclusions

This framework--this classification of moral prOposi-
tions into three levels--clarifies the issues in two ways.
First, it clears up several of the difficulties discussed
earlier in this chapter. Consider for instance the apparent
inconsistency between the kind of proposition to which "The
PhilOSOphy of Common Sense" refers, and those which make up
the Morality of Common Sense. Viewing appeals to Common Sense
as a procedure which has varying levels of subject matter
eliminates this charge of inconsistency. Similarly, the rela-
tionship between the third method Of verification and the body
Of Common Sense Morality is made clear by this interpretation.

Secondly, where this framework does not eliminate a

problem, it at least identifies more precisely just what the

109

problem is. Take, for example, the question Of whose consent
is to count. It turns out that while on every level universal
consent is desirable, nevertheless on each level the question
takes a slightly different form. On the third level, since
universal assent is unattainable for any precise statement,
the question becomes moot. On the first level, however, some-
thing close to universal assent actually exists, at least
implicitly,among those who weigh moral decisions. On the
second level the question retains its sharp edge, for it
becomes the question Of what is to count as sufficient reflec-
tion. The remaining problem identified early in this chapter
concerns the epistemic status of propositions which receive
general assent. Part of that problem is the tension which

Schultz identifies between what I have called CS and CS .

2 3
That issue is not resolved here, but it is sharpened. The

issue now is this: can the truth of CS prOpositions be

2
established without assuming the truth of CS3 prOpositions?
Is there an irreversible epistemological order among the
three levels? .

The issue of epistemic status raises another, equally
fundamental, question. Sidgwick, at every point, calls the
certainty attained by this procedure ”provisional." In a
sense, he, like Reid before him, ignores the question of

justification; but by using the word "provisional" he impli-

citly admits the difficulties. Why does universal assent

110

substantiate the truth of a prOposition? This central
question is left unconsidered.

Perhaps it is better answered obliquely. Like the
other remaining questions--How much reflection is sufficient?,
and, Is there a non-circular justification for the proposi-
tions of CSz?--it depends upon Sidgwick's concept Of ration-

ality. This is the topic to be addressed in the next chapter.
V. Intuition and Common Sense

The purpose of this chapter is to shed light on Sidg-
wick's concept Of intuition through an examination of the
closely allied tOpics, general assent and Common Sense. In
the light of the framework offered and discussed in the Section
IV, that question can now be asked more precisely. From the
point of view Of content, are the propositions of any of these
three levels those which Sidgwick would call intuitively cer-
tain, or self-evident? And, from the point of view of pro-
cedure, in what way does general assent add certainty to
propositions whose self-evidence has been determined in other
ways?

Let us look again at the prOpositions of each level.
The acceptance Of third level judgments, about individual
actions or classes of actions, is in fact called "intuition-
ism" by Sidgwick. This label, however, is misleading. What

Sidgwick means by "Intuitionism" is the belief that

111

we have intuitive 'knowledge' of the rightness Of

actions . . . ascertained by simply 'looking at'

the actions themselves1 without considering their

ulterior consequences.
This is, of course, a belief which Sidgwick rejects. His
examination of CS3 occurs in Book III of Methods; the title of
that book is "Intuitionism," and it concludes that such a view
of morality does not satisfy the requirements of rationality.
When Sidgwick talks about the intuitional foundation of ethics,
therefore, he is not referring to prOpositions Of the third
level.

Since Sidgwick calls the axioms of the second level
self-evident, and "self-evident" is part of what is meant by
"intuitively evident," it is clear that these axioms are part
of what Sidgwick means by intuition. What about propositions
Of the first level? He pays little attention to these, and
never labels them "intuition." Is there a difference between
the first and second levels which would justify that distinc-
tion? Propositions Of the first level are, after all, epis-
temologically prior to--more fundamental and more general
than-~those of the second; it would seem that if those Of the
second are known through intuition, so must be those of the
first.

There is no answer to this in Sidgwick's writing. The
bulk of it, Of course, concerns only propositions of the second

and third levels. Sidgwick notes his assumption (that there

 

1Ibid., p. 96.

112

is a knowable difference between right and wrong) and does
not return to it. We are left now, as we were at the end of
Section IV, with an important unsolved question: Should
Sidgwick's concept Of intuition be interpreted so as to in-
clude this fundamental first level ethical proposition? The
attempt in the next two chapters to deal with Sidgwick's con-
cept of rationality will be, in part, an attempt to deal with

the questions that have arisen here.

VI. Conclusion

 

This chapter began with the general question, Does
Sidgwick have a concept of self-evidence which is distinguish-
able from his concept of general assent? Since the whole
notion of general assent is linked to the terminology of
Scottish Common Sense PhilOSOphy, it was necessary to look
closely at what Sidgwick means by such terms. Problems were
found within his position--problems of whose consent is to
count, of unclear epistemic status, of confusion between con-
tent and method, of varying levels of generality in the pro-
positions cited; these problems were not solved by a close
examination of the historic source Of Common Sense PhilOSOphy,
the writings of Thomas Reid. In fact, that examination showed
that those problems are latent within Reid's work, and become
evident in Sidgwick's work in parge part because of his attempt
at greater precision. Finally, the chapter suggests classify-

ing the possible Objects of general assent by epistemological

113

level--that is, by degrees of generality and fundamentality.
Such a classification solves some of the problems mentioned
earlier, and gives greater focus to others. Although these
remaining questions concern propositions from all three levels,
each is particularly related to questions about the justifica-
tion for the axioms of C82. For example: Should propositions
of the first level be grouped with those of the second level,
as intuitively known? Is there a justification of CS which

2
does not depend upon the truth of CS ? How much reflection

3
is required for one's assent or dissent to count, in general
consensus tests of C82? The answer to these three questions
depends upon a close examination of the arguments with which
Sidgwick supports C82.

The even more fundamental remaining questions involve
the relationship between universal assent and epistemic status.
Does universal assent substantiate the truth of a proposition?
If so, how? It is this fundamental question which Sidgwick
(like Reid before him) ignores. The chapter began by asking
what Sidgwick's use of "Common Sense" could tell us about his
intuitionism. The answer must be that such appeals are, in
themselves, not shown by Sidgwick to be of value. They are at
most Of negative heuristic value. Are such appeals different
from appeals to self-evidence? 252mg.fggie, of course, the
answer is yes. But the preceding analysis shows that such

appeals are actually productive, for Sidgwick, only at the

first and second levels. The first level receives little

114

attention. SO, again, the "abstract transcendent axioms" Of
the second level are crucial. It is here that claims to self-
evidence and to consensus converge. The following chapter

will look at the arguments for these axioms more closely, with
the intention of delineating the concept of moral rationality
which underlies Sidgwick's appeals to self-evidence and to
consensus. This model of rationality may illuminate his intui-
tionism in a way which such categories as self-evidence and

consensus cannot.

CHAPTER IV

SIDGWICK AND INTUITIONISM

I. Introduction

 

Chapter II examined Sidgwick's notion Of self-evidence,
since he argues that if there is moral knowledge there must be
some self-evident foundation for it. Chapter III discussed
the role Of common sense in determining whether certain pro-
positions are self-evident. That chapter ends by distinguish-
ing various levels Of moral claims and suggesting that problems
about self-evidence differ at each Of the levels. The present
chapter, Chapter IV, looks at treatments of both third and
second level propositions and asks whether, for Sidgwick,
either Of these are possible instances Of intuitive knowledge.
The chapter distinguishes a number of senses of "intuition"
and asks whether Sidgwick is an intuitionist in any of these
senses. Along the way it considers various conflicting inter-
pretations of Sidgwick's methodology. It concludes that, for
Sidgwick, propositions of the third level are not known intui-
tively; propositions of the second level are possibly known

intuitively.

115

116

II. Various Meanings of "Intuition"
in Moral Theory

 

Sidgwick calls himself, and is always referred to as,
an intuitionist. The word "intuitionist," however, is Open
to a surprising number of interpretations, and therefore can
be misleading. In most cases the confusion arises from the
close connection in ethical theory between deontology and
intuitionism. This is an historical connection in that many
deontologists have been intuitionists. It is, for reasons I
will point to later, also a conceptual connection.

Sidgwick makes a number of distinctions among kinds
of intuitionism in moral theory, which are useful in avoiding
such a confusion. He distinguishes three kinds of "intuition-
ism," according to the generality of the rules to which
actions must conform. Perceptional Intuitionism maintains
that the agent can "see," in each act, its rightness or wrong-
ness. TO the extent that "rule" is apprOpriate here, the
word must mean a statement about the morality of this parti-
cular action at this time. (Sidgwick rejects Perceptual
Intuitionism without much argument.) The second kind of
intuitionism is Dogmatic Intuitionism, which holds as ulti-
mate "general rules prescribing different kinds of acts."1

This corresponds roughly with Common Sense Morality. Sidg-

wick rejects Dogmatic Intuitionism because the relationship

 

1Ibid., p. 103.

117

among the fundamental rules cannot be rationally explained.
This demand for a rational relationship among rules will be
dealt with in Section III of this chapter. The third kind is
Philosophical Intuitionism, which regards as ultimately valid

1 Sidgwick con-

"more universal and fundamental principles."
siders himself a philosophical intuitionist. In so labeling
himself, he is calling the abstract transcendent axioms of

the second level more authoritative than the Common Sense
Morality which comprises the third level.

However, Sidgwick himself notices the intertwining of
two distinct elements in his definition Of intuitionism, and
so distinguishes between a "narrow" and a "wide" sense Of the
word. In the narrower sense, intuitionism is the belief that
consequences alone dO not determine the morality of actions.
In the wide sense, "intuitionism" refers to the belief that
there are "self-evident principles relating to 'what ought to
be.”2 Sidgwick's wider sense, then, is an epistemological
one, and concerns the source of our moral knowledge. It is a
belief about what justifies claims to moral knowledge. Each
of the three kinds of intuitionism distinguished above--Per-
ceptional, Dogmatic, and Philosophical--is intuitionism in
this wider sense. It is this "wider" or epistemological sense

to which most Of this chapter is addressed.

 

1Ibid., p. 103 n.

2Ibid., p. 102 n.

118

In his "narrow" sense, intuitionism is a belief about
what justifies action. (In his initial definition he speaks
of the intuitionist as simply not considering the consequences
of actions; but he qualifies this almost immediately: "no
morality ever existed which did not consider ulterior conse-
quences to some extent."1 Thus the more careful statement
that consequences alone do not determine the rightness Of
actions.) This narrower sense refers to the content of the
intuitions, and may be called the deontological sense. Dog-
matic and Perceptional intuitionists belong in this deonto-
logical category, as well as in the "wider," epistemological
category. Philosophical intuitionism belongs only in the
epistemological category.

Sidgwick, who accepted a principle of utility, does
not consider himself an intuitionist in the narrower, deonto—
logical sense. Yet D. D. Raphael claims that Sidgwick is an
intuitionist in this sense. For the next few pages I will
review the various combinations Of epistemology and deontology
which occur in modern uses of the term "intuitionism." I will
also note, and reject, Raphael's arguments that Sidgwick may
rightly be called a deontologist (i.e., an intuitionist in

the narrow sense).

 

1Ibid., p. 96.

119
Some Modern Uses Of "Intuitionism" in Ethics

"Intuitionism," in current ethical writing, receives
a surprising variety of definitions. In many cases the
writer takes one element among the many which Sidgwick has
identified and makes that element the defining characteristic.
In Dogmatic Intuitionism, for instance, and in some
versions of PhiloSOphical Intuitionism, there is a plurality
of ultimate rules. John Rawls' use Of "intuitionism" centers
on that element:
I shall think of intuitionism in a more general way
than is customary: Namely, as the doctrine that t ere
is an irreducible family Of first principles . . .
In this sense of the word, uncommon though it is, Sidgwick is
an intuitionist. He maintains at least three fundamental
principles of morality: Prudence, Justice, and Benevolence.
Since these are second level principles, not third, Sidgwick
is an intuitionist.2
D. D. Raphael takes not one element from Sidgwick's
discussion, but two, confuses them, and as a result calls
Sidgwick a deontologist. He says at one point, ". . . Sidgwick

thinks that each of the methods of ethics depends on 'intuition,‘

 

lJohn Rawls, A Theory Of Justice (Cambridge, Mass.:
Belknap Press, 1971), p. 34.

2I have not argued here for the irreducibility of these
fundamental principles; this is usually assumed, but James
Newell explicitly argues for it. Newell, "Henry Sidgwick,"

p. l.

 

120

i.e., on at least one underivative principle of obligation
. . ."1 This is correct; here "intuition" is being used in
the wider, epistemological sense.

But he goes on to notice that the "names used by Sidg-
wick for the three methods are slightly confusing" since the
names Of the other two methods (Egoism and Utilitarianism)
"suggest that the principle of classification concerns the
character Of right action" rather than the source of the
knowledge about right action. Raphael therefore chooses to
see intuitionism "as referring to a belief about the character
of right action, and that only." At this point he has switched
to the narrower sense of intuition, as the belief that conse-
quences are not the sole determiner of moral rightness. He
proceeds immediately to confuse the two senses and to impute
to Sidgwick the view that "any properly ethical theory is in
part 'intuitional' (i.e., deontological)."2 He supports this
startling conclusion with the claim that

even teleological theories must rely on at least
one principle of obligation that does not itself
depend on the value of consequences . . . the
rightness of pursuing happiness itself is not a
matter of utility.

This view he takes to be Sidgwick's. What Raphael has done,

however, is confuse two issues. One concerns the justification

 

1David D. Raphael, "Sidgwick on Intuitionism," Monist
58 (July, 1974):409.

2Ibid., p. 407.

31hid.

121

of actions: to hold that consequences alone do not make an
action right is to hold a deontological position. The other
concerns the justification of belief. To hold that funda-
mental principles 22223 the rightness of actions carry their
own justification--are self-evident--is to be an ethical
intuitionist. The first issue is normative; the second is
epistemological. The first concerns intuitionism in Sidg-
wick's narrow sense; the second concerns intuitionism in Sidg-
wick's wider sense.

Raphael's claim is that a teleological belief about
the rightness of actions cannot itself be justified on teleo-
logical grounds; thus it must be a deontological belief. But
teleological beliefs are not justified on deontological grounds,
either. Only actions knuiintentions, policies, etc.) can be
justified on either teleological or deontological grounds;
beliefs must be justified on epistemological grounds.

Kurt Baier also argues for a connection between deon-
tology and epistemological intuitionism. The connection he
draws, however, does not depend upon a confusion; furthermore,
it helps to explain the strong historical coincidence between
the two. Baier argues that any deontologist must be an intui-
tionist. Since deontologists rely upon something besides the
consequences of an act, in judging whether or not that act is
right, they will (usually or always) "claim that rightness is

a simple property which we can 'see' or 'intuit' in an act

122

. . ."1 Thus the Dogmatic Intuitionist, who has a plurality

of third level principles, none of them justified in terms of
a second level teleological principle, is an intuitionist both
in the wide, epistemological sense and in the narrow, deonto-
logical sense. Most theorists called ethical intuitionists
are in fact both.

But here, too, one must be careful. Frankena appears
to take a single element from Sidgwick's discussion of intui-
tionism and make that the defining characteristic when he
writes that intuitionism holds that "Obligation is indefinable

2 This alone would not make one an intuition-

and non-natural."
ist in the wide sense, for this is a claim about the meaning
of principles of Obligation. The epistemological intuitionist
makes a stronger claim: the truth of such principles must be
self-evident.

The remainder Of this chapter deals with intuitionism
in the epistemological sense. It deals separately with the
propositions of the second and third level, asking for those

of each level whether it is apprOpriate to call Sidgwick an

intuitionist in this sense.

 

1Kurt Baier, "The Moral Point of View," in Readings
in Ethical Theory, ed. Wilfred Sellars and John HOSpers (New
York: Appleton-Century-Crofts, p. 1970), p. 471.

2William Frankena, "Obligation and Motivation in
Recent Moral Philosophy," in Hospers and Sellars, Readings
in Ethical Theopy, p. 709.

 

 

 

 

123

III. Intuition and Propositions
of the Third Level

 

Propositions Of the third level concern particular
actions and particular kinds of actions. What Sidgwick calls
Common Sense Morality is made up of prOpositions of this kind.
This Section asks whether Sidgwick claims that we know these
propositions in some special, direct way that should be called
intuition. The discussion begins with a description of a
technique used by John Rawls, called "reflective equilibrium."
It has been argued that use of this technique commits the user
to either subjectivism or to intuitionism (for judgments of
both second and third level). Sidgwick claims that a relation-
ship Of "dependent validity" holds between third level and
second level propositions. This relationship differs from
reflective equilibrium (as that phrase is commonly understood),
and does not commit Sidgwick to intuitionism for third level
propositions.

"Reflective Equilibrium" and
Epistemological Intuitionism

"Reflective equilibrium" is a term introduced by John

1

Rawls in A Theory of Justice. It describes a technique used

 

in a great deal of contemporary ethical work as well as in
Rawls' own. It has been claimed that Sidgwick uses that tech-

nique. After describing that technique I will consider whether

 

lJohn Rawls, A Theory of Justice, p. 20.

 

124

its use commits the user to epistemological intuitionism. I
will then consider whether Sidgwick in fact uses that tech-
nique.

Rawls begins with moral convictions about particular
actions and particular kinds of actions (Rawls frequently
uses the term "intuitions" for such judgments; but I will
refer to them as judgments of the third level). When these
convictions result from "impartial judgment not likely to be
distorted by an excessive attention to our own interests,"

1 which our

such convictions "are provisional fixed points"
general principles (second level) should fit. When, however,
we are not certain about a third level moral decision, then
our second level general principles should be used tO settle
our uncertainty. These second level principles derive their
authority from their capacity "to accommodate our firmest con-
victions and to provide guidance where guidance is needed."2
There is a kind of balance, then, between two levels, each
deriving its authority from its relation to the other, and
each subject to revision when the other changes. This dynamic
balance is what Rawls calls "reflective equilibrium."

This technique has been criticized. Because of it,
Rawls has been called an intuitionist by those for whom that

term is pejorative, who consider intuitionism a retreat to a

non-rational and ultimately subjective position. A

 

Ibid.

Ibid.

125

particularly vitriolic attack is made by R. M. Hare. As he
interprets it, the technique gives ultimate authority to the
third level judgments of Rawls and his readers. His criti-
cism is worth quoting at some length.

To speak like this . . . is to make the truth of
the theory depend on agreement with peOple's Opin-
ions . . . it is depressing to find a self-styled
objectivist falling as deeply into [subjectivism]
as Rawls does--depressing, because it makes one feel
that this essentially simple distinction will never
be understood: the distinction between the view
that thinking can make it so (which is in general
false) and the view that if we are to say something
sincerely, we must be able to accept it (which is a
tautology).l

h

On Hare's view, then, calling third level judgments authorita-
tive is either an empty claim or a false one; and to persist
in this is to remove any grounds for claiming Objectivity in
ethics.
A more moderate form of this criticism is given by

David Lyons, who considers the technique (Lyons calls it a
"coherence" theory) inevitable but necessarily imperfect.
His criticism contains two major points. First, he echoes
Hare's concern about subjectivism. In Lyon's words,

the justificatory force of coherence arguments is

unclear. Pure coherence arguments seem to move us

in a circle, between our current attitudes and the

principles they supposedly manifest . . . we can
still wonder whether [our basic convictions] express

 

1R. M. Hare, "Rawls' Theory Of Justice," in Reading
Rawls, ed. Norman Daniels (New York: Basic Books, Inc.,
1976), p. 83.

126

any more than arbitrary commitments or sentiments
that we happen now to share.1

Lyons considers a defense of the Objectivity of these third
level convictions. It might be argued, he wrotes, that

coherence arguments really do justify moral princi-
ples because they provide the strongest arguments
that are possible for them--and this constitutes
justification, properly understood.

Such a position, however, assumes

a fallacious principle that something is justified 3
when all possible arguments for it have been given.

It also leads to an absurdity, for it "implies the possibility
of justifying unjustifiable assertions."

Secondly, he points out an internal problem in any
coherence theory Of morality. Even on its own terms, the
justification it provides is at best provisional.

[N]o coherence argument can be conclusive. For none
of our moral judgments . . . can be considered incor-
rigible. And it is, moreover, unlikely that any set
Of considered judgments will mesh neatly with an
initially proposed, intuitively attractive and illu-
minating set of principles that might seem to ground
them and to extend them in an acceptable way. To
achieve a satisfactory fit between judgments and
principles, some judgments must be diicounted, initial
principles must be modified, or both.

In addition, our judgments concern only those situations which

we have had occasion to consider--we can never eliminate the

 

lDavid Lyons, "Nature and Soundness Of the Contract
and Coherence Arguments," in Daniels, ed., Reading Rawls, p. 146.

2

 

Ibid., p. 147.

3Ibid.

4Ibid., p. 146.

127

possibility of new situations arising. Therefore "Rawls does
not and cannot claim that his principles match our considered
judgments perfectly."1

It is likely, then, that users of reflective equili-
brium must choose between subjectivism and intuitionism.
Either the whole balanced set of judgments has no Objective
foundation, or all of the judgments (third level and second
level) have some kind of intuitive authority. (As mentioned
above, in Section II, and in the treatment of Hare, there are
many who see those alternatives as amounting to the same
thing--i.e., it is hard to establish the objectivity of sup-
posed intuitions.) And if they do classify third level judg-
ments as intuitive knowledge, they run into the problems
mentioned by Lyons in making these judgments cohere with one
another and with the second level principles.

Since the question at issue here is whether or not
Sidgwick is an intuitionist with respect to third level judg-
ments, it becomes important to consider whether or not Sidg-

wick does in fact use the technique described above as reflec-

tive equilibrium.
Sidgwick and Reflective Equilibrium

The resemblance between Rawls and Sidgwick in this

respect is strong. However, if Sidgwick is using any form of

 

Ibid 0

128

reflective equilibrium, it is not one which commits him to
epistemological intuitionism for propositions of the third
level.

To substantiate that claim, the next few pages will
present a detailed discussion of the relationship which Sidg-
wick claims exists between second and third level propositions.
That relationship is complex and easily misunderstood. An
example of that misunderstanding appears in Robert Schultz's
argument that Sidgwick's concept Of proof in ethics is essen-
tially incoherent. Schultz's conclusion-~that Sidgwick oscil-
lates between normative and descriptive concepts of proof--
has already been discussed, and refutation attempted (pp. 64-
66 above). However, I will review the reasons which lead
him to that conclusion, for those reasons involve a misinter-
pretation Of the crucial relationship between Sidgwick's
second and third level prOpositions. Schultz claims that
Sidgwick uses the less certain third level propositions to
prove the more certain second level propositions.

Schultz refers to Sidgwick's claim that third level
judgments have "general validity" which is only a "dependent
and subordinate validity."1 To Schultz, this notion of "sub-
ordinate validity" is incoherent. The relevant passage in
Methods occurs when Sidgwick is arguing for the superior valid-

ity of Utilitarianism. Common Sense Morality has

 

1Sidgwick, Methods, pp. 421-422.

129

only a dependent and subordinate validity . . .
Utilitarianism sustains the general validity of

the current moral judgments . . . and at the same
time affords a principle of synthesis, and a method
for binding the unconnected and occasionally con-
flicting principles of common moril reasoning into
a complete and harmonious system.

To summarize his Objection, Schultz says
[T]he question remains . . . how a principle that
is, in any sense, 'argued for' or 'proved from'
premises that are the rules Of common sense moral-
ity, c n . . . be used to correct and extend those
rules.
In answering Schultz's charges--and in substantiating
the claim that Sidgwick is not an epistemological intuitionist
for propositions of the third level--I will be indebted to

Schneewind's interpretation Of the central argument in Methods

of Ethics.3 Schneewind points out the importance of the his-

 

torical context in understanding Methods Of Ethics.

[C]ommon sense morality had come by the middle of

the 1860's to occupy a central place in . . . ethical
controversies . . . the problem with which Sidgwick
is mainly concerned [is]: to what extent can common
sense morality be systematized on a self-evident
basis?

 

1Ibid.

2Schultz, "Proof," p. 350.

3A5 Schneewind points out, the general outline Of
Sidgwick's argument in Methods of Ethics is obscure. C. D.
Broad treats the book as a collectiOn Of valuable reflections,
and points out no overriding argument at all. Many readers
have taken it to be a defense of utilitarianism, and yet Sidg-
wick himself denies this.

4Schneewind, Sidgwick, p. 193.

130

He identifies two major components in Sidgwick's answer to
this question; Schneewind calls them the systematization and
dependence arguments. Each requires detailed explication.
I will begin with the dependence argument.

"The dependence argument is Sidgwick's attempt to show
. . . that reason requires first principles of a kind which

common sense does not supply."1

Sidgwick distinguishes between
the ultimately arbitrary facts of physical nature and the
(ideally) rational facts Of moral value. In the first realm
we will, in the second realm we will not, admit variations
for which we can discover no rational explanation. But all
that Common Sense Morality gives us is a list of actions,
some of which are right and others wrong, with no explanation
for the difference between them.

The contrast with science is less helpful than either
Sidgwick or Schneewind thought; Sidgwick's underlying point
is an ideal of rational explanation. Schneewind refers us tO
Lectures on the Philosophy Of Kant, but there Sidgwick dis-
cusses such an ideal for science too:

For the purposes Of science, the judgment that some
members of a class have an attribute . . . immediately
suggests the question, By what other characteristics
is this portion of the class distinguished? We then
seek at once to turn the 'some' into the 'all' of a

sub class, by ascertaining their common characteris-
tics.

 

1Ibid., p. 280.

2Sidgwick, Kant, p. 84.

131

Sidgwick describes the relationship between Common
Sense Morality and Utilitarianism as twofold: "at once nega—
tive and positive."1 This phrase occurs in Book IV, Chapter
II, "The Proof Of Utilitarianism." Sidgwick takes this
Opportunity to relate the conclusions of Book III (his examina-
tion of Common Sense Morality) to his project in Book IV (the
proof of utilitarianism). It is here that Schneewind's recon-
struction is important and revealing. What Sidgwick calls
the "negative" stage, Schneewind calls the "dependence" argu-
ment; what Sidgwick leaves implicit, Schneewind makes

explicit.2

The "dependence" argument is so named (by Schnee-
wind) because its conclusion is that Common Sense Morality has
only a "dependent and subordinate validity."3 As he sees it,
the dependence argument begins after Sidgwick has established
that the principles Of Common Sense Morality are not self—
evident; that is, the principles do not survive scrutiny by
the three methods of verification described in Chapter II Of
this dissertation. The first method of verification offered
means of evaluating claims to self-evidence; the second
demanded mutual consistency among a set Of principles; the

third demanded general agreement. When Common Sense Morality

is examined carefully, it offers generally agreed upon and

 

lSidgwick, Methods, p. 421.

2Schneewind, Sidggick, pp. 264-265.

 

3Sidgwick, Methods, p. 422.

132

mutually consistent principles only at the cost of
self-evidence.l
This is Sidgwick's conclusion in Methods after 150
pages examining Common Sense Morality. At this point Schnee-
wind points to a discussion outside of Methods which describes
more fully what might count as self-evidence in ethical prin-

ciples.

In Lectures on the PhilOSOphy_Of Kant,2 Sidgwick dis-

 

cusses what "independent" validity would mean. Such principles
would, for one thing, be "absolutely valid": each "conclu-
sively determines the actual moral status"3 of acts which fall
under it. In addition, the principles are ultimate: "there

is no valid principle of greater generality . . ."4 The
dependence argument is Sidgwick's attempt to show that the
results of the examination of common sense morality are incom-
patible with these two related claims . . ."5 Whenever common
sense morality satisfies one of these conditions it fails the
other. If the rule is "ultimate"--not dependent on a rule of

greater generality--it is tOO vague to apply to actual cases.

If it is sufficiently precise to apply to actual cases, then

 

lSidgwick, Methods, pp. 342-343.

2Sidgwick, Kant, p. 2.

3Schneewind, Sidgwick, p. 279.

 

4Ibid.

51bid., p. 280.

133

the statement of it includes exceptions and limits, and such
statements and limits, if rational--i.e., not merely arbitrary
--must be determined by some other principle.

If morality is rational, in short, the rules taken

by dogmatic intuitionism as first principles can

only determine putative rightness. But first prin-

ciples, it is agreed, must determine actual rightness,

since otherwise the rationality of morality is incom-

plete. The dependence argument shows, then, that from

the main result of the examination of common-sense

morality, it follows that the principles the intui-

tionist holds have only a 'dependent and subordinate

validity.‘l

The positive side of the argument--the systematization

argument--is the claim that utilitarianism supplies the ratiOn-
ality which is missing in common sense morality. The system-
atization argument includes an account Of what it means for
some principle to be "superior in validity" to others: it
will be "at least as wide-ranging in its scope" as the deriva-
tive principles; it "determines the limits and exceptions of
other principles, while none determines such restrictions to
it." In morality, therefore, such a principle will determine
the actual rightness of actions. It is then Sidgwick's major
burden to show that Utilitarianism does have this relationship.

to the morality Of common sense. That effort is discussed in

Chapter V.

 

l
p. 422.

Schneewind, p. 282. His reference is to Methods,

134

Sidgwick on Third Level PrOpositions and
Epistemological Intuitionism

Whether or not this systematization and dependency
relationship is to be called reflective equilibrium is not
important; nor is the issue whether or not Rawls himself uses
Sidgwick's technique. What is important is whether Sidgwick's
position commits him to the dilemma which seems to arise from
some kinds of reflective equilibrium: that is, either subjec-
tivism or an intuitionism which includes intuitionism for
third level propositions.

Peter Singer argues that it does not. His thinking is
interesting because, although I think his conclusion is cor-
rect, the argument he presents to support it is inadequate.

As Singer sees it, Sidgwick uses the data Of Common Sense
Morality
not as a means of establishing the validity of utili-
tarianism, but as a means Of winning over . . . one
who happens to hold a particular moral position,
Opposed to utilitarianism--namely, the philosophy of
common sense morality.
Singer calls this an 3g hominem argument. There is no doubt
that this is in part what Sidgwick is doing, for he says so
frequently. The question is whether this is the only use he
makes of third level moral judgments, for if it is, then Sidg-

wick is clearly not granting that these judgments constitute

intuitive knowledge. In Singer's attempt to establish that

 

1Peter Singer, "Sidgwick and Reflective Equilibrium,"
Monist 58 (July, 1974):499.

135

Sidgwick uses Common Sense Morality only in this 3g hominem
way, Singer makes several additional points. He mentions
Sidgwick's insistence that the Oecumenical Test gives us only
provisional certainty. This alone will not establish Singer's
point, because Sidgwick attaches the same qualification to all
his Methods of Verification. Whatever certainty can be reached
through Intuitive Verification-~regarding the second level
axioms--will also be provisional. This qualification attached
to third level judgments, therefore, does not prove that Sidg-
wick considers them less authoritative than second level judg-
ments.

TO support his claim Singer also points out that the
subject matter of Sidgwick's inquiry is the method of Common
Sense Morality rather than its conclusions; in Reflective

Equilibrium it is the conclusions which serve as "fixed points"

 

and must be systematized. However, the method Of Common Sense
Morality is appeal tO a set of rules about action. Rules in
this sense could be the subject of Reflective Equilibrium,
just as conclusions about particular acts can be.

Singer's final argument is more effective, it relies
on Sidgwick's acceptance of egoism as ultimately rational.
Egoism is incompatible with Common Sense Morality, but this
incompatibility in no way counts against Sidgwick's conviction
Of its rationality. Since Rational Egoism is a second level

proposition, this argument offers real support for the claim

136

that Sidgwick is not an epistemological intuitionist for
prOpositions Of the third level.

The most convincing support for that conclusion, how-
ever, lies in a close look at the systematization and depen-
dency arguments themselves. The crucial question is the
correct meaning to be applied to the phrase "dependent validity."
Sidgwick uses the phrase, as Singer claims, as part of an 3g
hominem argument; but only as that? As I have interpreted
Schneewind's interpretation, the "dependent" in the phrase
"dependent validity" means at most "derivative." Third level
judgments are indeed knowledge claims, but they are not claims
to intuitive knowledge. If anything is to count as intuitive
knowledge it is the second level axioms; the third level judg-
ments count as knowledge only because of their relationship
to the second level axioms.

Perhaps it is Sidgwick's characteristic caution which
leads to an Opposing interpretation, for he insists upon the
provisional nature of our certainty even at the second level,
and when these axioms lead to conclusions which contradict
Common Sense Morality we are told to be wary. In such a case,
something must be mistaken. But no matter how firm our third
level conviction, Sidgwick does not assume that the second level
axiom must be reformed to fit it.

If, then, Sidgwick is not using Reflective Equilibrium
in spite of the strong resemblances between his technique and

Rawls', then the arguments which lead Hare to call Rawls an

137

intuitionist cannot be used against Sidgwick. And if third
level convictions are helpful guards against error rather than
authoritative--if their validity is only derivative from that
of the second level axioms--then there remains no reason for
calling Sidgwick an intuitionist in respect to third level
propositions. He does not claim a special, direct knowledge

with certainty of those moral truths.

IV. Intuition and Propositions Of the Second Level

 

A more significant question is whether Sidgwick's
treatment of the second level prOpositions is apprOpriately
called epistemological intuitionism. It is, after all, of
the self-evident axioms that Sidgwick was Speaking when he
called himself a philOSOphical intuitionist; and the term is
meant to contrast with the Perceptional and Dogmatic intui-
tionism which he finds wanting.

In order to discuss this issue it will be necessary to
return to a consideration of just what epistemological intui-
tionism is. The second section of this chapter has discussed
the frequent linking of intuitionism and deontology, pointed
out the reasons for this, and attempted to distinguish intui-
tionism as a view about the content of morality from intuition-
ism as a view about the source of moral knowledge. The latter
is the most common use of the term; it can be distinguished

from other (usually mixed) uses by calling it epistemological

138

intuitionism. For the rest of this chapter I will continue
to use the term intuitionism in a strictly epistemological
sense.
Epistemological Intuitionism:
Further Examination

At this point the previously unexplicated concept of
"epistemological intuitionism" needs to be clarified. Intui-
tionism seems most attractive on a foundationalist approach
to knowledge. Chapter I, examining the elements in Sidgwick's
concept of morality which lead him to intuitionism, lists as
a crucial thesis, "All knowledge is either self-evident or
derived from self-evident premises." This dissertation neither
examines nor evaluates that premise. Nevertheless its impor-
tance, and the fact that it can be questioned, must be noted.

In such a foundationalist framework, intuitionism
seems necessary to avoid infinite regress. We must arrive at
a non-inferential base and, if we are to avoid skepticism,
that non-inferential base must count as knowledge. But what--
if anything--can justify such basic beliefs? What can serve
to distinguish truth from falsity?

It is helpful to use Richard Rorty's two senses Of

intuition: a weak sense and a strong sense.1 In what might

 

1Encyclopedia of Philosophy, 5. v. "Intuition," by
Richard Rorty.

139

be called the weak sense, intuitive knowledge is synonymous
with non-inferential knowledge.
Weak sense: S knows the p intuitively only if
a) p is true
b) S is justified in believing that p and
c) S's knowledge is non-inferential.
This kind of intuition can be known without supporting argu-
ments; on the other hand, should doubts arise about the truth
Of the proposition, such support can be helpful in settling
the doubts. A paradigmatic instance of intuition in this
weak sense is our knowledge that we have two legs. We do
not know this inferentially, we know it directly. Yet, should
doubts arise about the truth of the proposition, we can look,
count, ask other peOple, and so on. Information from these
other sources can legitimately allow us to conclude that the
proposition in question is true. Thus intuition in the weak
sense refers to propositions which are capable of being knoWn
directly egg capable of being known indirectly.
In the stronger sense, intuitive knowledge is a parti-
cular form of non-inferential knowledge.
Strong sense: S knows that p intuitively only if
a) p is true
b) S is justified in believing that p and
c) there are no accepted procedures for resolving

doubts about the truth Of p.

140

Propositions which are known intuitively in the
strong sense form a subset of those known intuitively in
the weak sense. There are two paradigmatic instances Of
intuition in the strong sense: "first person statements
about those psychological states to which one has 'privileged
access' and underived a priori truths."1

To a non-foundationalist, knowledge in neither of
these two senses should be called intuitive knowledge. Tom
Regan, for instance, argues that any claim to moral intuition
must fail. He, too, distinguishes two kinds of supposedly
intuitive knowledge; although his terminology differs, his
dilemma can be phrased in terms of what is called above the
strong and weak senses of intuition. Either there is no way
of confirming our possession of intuitive moral knowledge
(this is the strong sense identified above); or there is such
a method (this is the weak sense). But any such method Of
confirmation "would open up the possibility that p could be
known gggintuitively . . ." and thus destroy any necessity
for claiming the knowledge is gained intuitively. Since this
necessity is the only factor making intuitionism attractive,

the existence of a decision procedure for so-called intuitions

 

Ibid.

141

would eliminate the most important support for claiming that
we have intuitive knowledge in this sense.1

But, to Regan, an equally serious difficulty befalls
any claim to intuitive knowledge in the strong sense. He
paraphrases a typical defense of ethical intuitionism as
follows:

[T]here are ends which every rational man must agree

are worthwhile . . . (it is part of the meaning Of

"a rational man" that he do so); . . . we know which

ends these are.2
This is Regan's paraphrase Of an intuitionist's position.
Regan concludes that this position is equivalent to saying,
"I know that I know, and I can say nothing about it." To him,
this is an unjustified claim to knowledge, and therefore should
not be called a claim to knowledge at all. He chooses instead
to grant the claims made above about rational ends, and yet
deny that these constitute knowledge.

Having noted this non-foundationalist attack on the
coherence of "intuitive knowledge" I will simply repeat that
Sidgwick is a foundationalist. If we could establish that
there are "ends which every rational man must agree are worth-
while" Sidgwick would call this knowledge. The acceptance of

intuition in the strong sense, therefore, is attractive for

all of the reasons which make foundationalism attractive (and

 

lTom Regan, "McCloskey on Rational Ends: The Dilemma
of Intuitionism," Journal of Value Inquiry 9 (Spring, 1975):
48-58.

 

2Ibid., p. 54.

142

the reasons which make foundationalism attractive are outside
the scope of this dissertation).

Schneewind summarizes the history of ethics in the
centuries preceding Sidgwick. It is a history of two compet-
ing foundationalist approaches to morality. One, which he
calls a Reidian notion, emphasizes the authoritativeness of
the moral faculty rather than its rationality, and explicitly
compares the moral sense with the other senses.1 The other,
which he calls a Kantian notion, identifies the moral faculty
with reason. Schneewind finds an unrecognized tension between
these two concepts in the intuitional ethics of William Whewell,
who

does occasionally slip into treating intuition as a

power . . . Of simply seeing some truth . . . But

more commonly Whewell attempts to avoid any untest-

able appeal to a cognizing faculty and to rest

morality on propositions which are 'seen' because

they plainly display their own inherent rationality.
In terms of this chapter, the Reidian notion is one example Of
intuition in the "strong" sense of the word (where S knows p
intuitively only if p is true, S's belief is justified, and
no decision procedure exists). What Schneewind calls the
Kantian sense includes at least the "weak" sense of intuition
(in which S knows p intuitively only if p is true, S's belief

is justified, and 8's knowledge is non-inferential).

 

lSchneewind, Sidgwick, pp. 67—68, 72, 106-107.

2ihid., p. 113.

143

Now Schneewind also maintains that, although "Sidgwick
makes few explicit comments on Whewell, much of the Methods

is an implicit critique Of his work."1

And Schneewind places
Sidgwick in the Kantian tradition.

Clearly Sidgwick ig seeking a foundation for ethics
which falls within that tradition. He calls the second level
axioms self-evident (or, apparently self-evident), and he
speaks Of these as having greater validity than the third
level judgments. He is seeking principles which, in Schnee-
wind's words, "exhibit their inherent rationality." As Sidg-
wick put it, these principles have "superior validity." If
the ggly_kind of intuition which "exhibits its inherent
rationality" is intuition in the weak sense, then Sidgwick is
seeking a foundation for ethics which is intuitive in the weak
sense only. But if there is room for both a strong and a weak
sense of intuition within the Kantian tradition, further ques-
tions arise. Is it possible for a principle which "exhibits
its own rationality" to be a principle for which there is no
decision procedure? This seems unlikely, since the very
meaning of rationality includes an appeal to shared criteria
by which to separate the rational from the irrational. Sup-
pose, however, that one appeals to a standard "which all
rational people accept" but one is unable to explicate yhy

all rational people accept that standard. And, should someone

 

1Ibid., p. 112.

144

speak rationally and yet reject the appeal, one is unable to
give reasons which would change the rejector's mind. Such

an appeal is to intuition in the strong sense; but the appeal
belongs within the Kantian tradition.

Granting Schneewind's claim, then, that Sidgwick
belongs within the Kantian tradition, we are nevertheless left
with a significant question. Does he treat second level
propositions as possible instances of intuition in the strong
sense, or only in the weak sense? (Those known in the strong
sense form a subset of those known in the weak sense.)

The possible parallel between underived a priori truths
(paradigmatic instances of intuition in the strong sense) and
second level axioms is one major concern. Sidgwick's claims
about good are similar in epistemic status to these second
level axioms, and are also claims to intuitive knowledge.

The statements about good will be examined first, followed by
an examination of statements about ought (the second level
axioms). In both cases, it will be seen that Sidgwick seeks
foundations which are intuitive in the weak sense. In the
case Of "good" statements, he must settle for claims to intui-
tive knowledge in the strong sense. In the case of "ought"
statements, his success in attaining foundations which are
intuitive in the strong sense depends upon the success of the

considerations with which he supports these foundations.

145
Hedonism and Intuitionism

Sidgwick writes that the only ultimate good is the
pleasure of sentient beings. He bases this claim on an
explication Of "good" as "what one would desire if one's
desires were in harmony with reason . . ."1 Various other
explanations of "good" are rejected because they lack any
normative component thitoday's vocabulary; in Sidgwick's own,
they lack appeal to a "universally valid standard." See the
discussion in Chapter I of this dissertation). That normative
component is reason, and thus his own explication includes
the phrase, "With reason."

Sidgwick then is making two centrally significant
claims about good. First, he says that to call something
good is to claim that it is what would be desired if one's
desires were rational; secondly, that what in fact is ulti-
mately good--what in fact is the ultimate Object of rational
desire--is the pleasure of sentient beings. I will argue
that for Sidgwick each of those claims is a claim to intuitive
knowledge in the strong sense.

Sidgwick's discussion actually has two steps; in one
he discusses the meaning of "my good on the whole,” and in
the next the meaning of "ultimate good on the whole." These

distinctions are crucial in the discussion in Chapter V

 

lSidgwick, Methods, p. 112.

146

concerning the Dualism Of Practical Reason, but the present
discussion is confined to the elements common to both con-
cepts Of good.

He argues that "good" cannot be reduced to "what is
desired" or even to what would be desired if its consequences
were "accurately foreseen and adequately realized in imagina-
tion." He rejects these definitions because the notion of
good (strictly speaking, he is dealing here with "my good on

the whole") is "authoritative: and therefore carries with it

 

implicitly a rational dictate to aim at this end." And so
his final position is that "good" ("ultimate good on the
whole") means "what as a rational being I should desire and
seek to realize, assuming myself to have an equal concern for
311 existence." Part Of what it means to be rational is sup-
plied by the final phrase about concern for all existence.
What is good is the object of rational desire.l

He argues for this position by discussing the various
ways in which we use the word "good." He considers various
possible definitions and rejects them, because they do not
adequately convey the meaning Of the word. If this discussion
of definition is an appeal to intuition at all, it is to
intuition in the weak sense. There is a decision procedure,
although not a precise one; that procedure is reference to

our common understanding Of what a particular word means.

 

Ibid.

147

Let us turn, then, to his second claim about "good."
Only the pleasure of sentient beings is in fact good; only
the pleasure of sentient beings can be the object of rational
desire.

Sidgwick's argument begins with a consideration of
various things which might be called good: virtues and excel-
lence in particular.

However immediately the excellent quality of such

gifts and skills may be recognized and admired,

reflection shows that they are only valuable on

account of the good or desirable conscious life

in which they are or will be actualized . . .
The issue, then, is what is to count as desirable conscious-
ness. Again he considers alternatives; pleasure is clearly
desirable, but is it the only ultimately desirable quality of
consciousness? Knowledge Of the truth and the contemplation
of beauty, for instance, seem also to be ultimately desirable.
Sidgwick thinks they are not, and asks the reader to follow a
familiar procedure.

I appeal firstly to his intuitive judgment after

due consideration of the question . . . and secondly

to a comprehensive comparison Of the ordinary judg-

ments Of mankind . . . when (to use Butler's phrase)

we 'sit down in a cool hour,’ we can only justify to

ourselves the importance that we attach to any of

these [alternative objects of desire] by considering

its con uciveness . . . to the happiness of sentient
beings.

 

1Ibid., p. 395.

2Ibid., p. 401.

148

Sidgwick appeals to "reflection" and to "intuitive
judgment." This appeal is a claim to intuitive knowledge
in the strong sense. If someone were to sit down "in a cool
hour" and decide that something besides pleasurable conscious-
ness is desirable for its own sake, Sidgwick could say nothing
but "think again."1

PeOple do, upon reflection, differ about what is of
ultimate value. These differences are both more common and
of more practical effect than differences about the paradig-
matic cases of.intuition in the strong sense. Because of
these significant disagreements, intuition in the strong
sense is a more problematic basis for ethics than for empiri-
cal knowledge.

Sidgwick's discussion Of ultimate good invokes the
"ordinary judgments of mankind." Using the distinctions pro-
vided by this chapter, we can now understand more fully Sidg-
wick's frequent use of appeals to consensus in support of
"apparent self-evidence." The relationship between consensus
and self-evidence was an unanswered question at the end of
Chapter III. Consensus counts in support of the self-evidence

of a claim for Sidgwick, because he is constantly seeking

ethical foundations which are intuitive in the weak sense

 

1That is, if the competing candidate for ultimate
good were a fairly reasonable one. The next chapter of this
dissertation summarizes Sidgwick's rejection of such candi-
dates as fame. He believes he could persuade someone that
fame is only Of instrumental value.

149

only. Since the weak sense of intuition is simply
non-inferential knowledge for which convincing proof is
available should it be needed, the common Opinions Of man-
kind would be likely to agree upon such claims. But in the
case, for instance, of what is to count as ultimate good he
actually must make do with an appeal to intuition in the
strong sense. Consensus helps to support the truth of such
claims because it indicates the possible existence of some
shared but unexplicated understanding that accepting this
claim is reasonable. If these shared but unexplicated rea-
sons could only be made explicit, we could move from the
strong sense of intuition to the weak.

Finally, a kind Of red herring must be considered.
One of the paradigmatic instances Of intuitive knowledge in
the strong sense is privileged first person reports. G. E.
Moore, who writes that goodness is an intuited non-natural
property,is probably claiming a parallel between the privi-
1eged first person reports and this recognition Of a non-
natural prOperty. Moore also acknowledges a debt to Sidgwick,
and so it is tempting to think that Sidgwick too claims such
a parallel.

Moore praises Sidgwick's avoidance (unique, Moore

thought, among ethical writers) of the naturalistic fallacy.l

 

1G. E. Moore, Principia Ethica (Cambridge: Cambridge
University Press, 1903), p. 17.

 

150

Moore also distinguishes two kinds of intuitionism: that
whose subject matter is "The Good" (Moore: "What kind of
things ought to exist for their own sakes?") and those con-
cerning rightness or Obligation (Moore: "What kind Of actions
ought we to perform?"). He is an intuitionist only about the
first class of proposition, and even there he is careful to
limit the meaning of that term.

[Wlhen I call such propositions 'Intuitions,' I mean

merely that any proposition whatever is true, because

we cognise it in a particular way . . . [I]n every

way in which it is possible to cognise a true prOposi-

tion, it is also pOSSible to cognize a false one.
Earlier in this passage Moore seems to imply that Sidgwick,
too, properly understood, is an intuitionist in this sense,
and only in this sense. Sidgwick does resemble Moore in cau-
tion; that is, Sidgwick always speaks of apparent self-evidence
and never claims complete certainty. Sidgwick also resembles
Moore in claiming that ethical statements cannot be reduced
to "natural" ones.

But if we ask the further question-~does Sidgwick's
argument parallel such claims as that each of us knows imme-
diately and incorrigibly whether s/he is in pain--then the
answer must be no. Surely the question of what each of us
finds desirable can only be answered through a privileged

first person report. But Sidgwick insistently points out that

there is a normative element in the concept Of Good, and this

 

1Ibid., p. x.

151

cannot be reduced to physical or psychological statements.
The "non-natural" component in Sidgwick's explanation of
"goodness" is not a property but an appeal to rationality.
Finally, Moore is incorrect if he means that Sidgwick
is not an intuitionist about "what kind of actions we ought
to perform." The claim that Sidgwick is an intuitionist in
the strong sense about the second level axioms constitutes

the final sections of the present chapter.
The Fundamental Axioms and Intuition

Is Sidgwick's treatment of the second level axioms
one that implies intuitionism in the strong sense? One way
to approach that question is to compare these axioms with one
of the paradigmatic instances: our knowledge of underived a
priori claims which underlie scientific knowledge. These a
priori claims include, for instance, the principles of causal-
ity and of the uniformity of nature. The context in which
they are made is a foundationalist account of empirical
knowledge, just as Sidgwick is working within a foundational-
ist account of ethical knowledge.

Some parallels between the ethical axioms and the
scientific axioms are Obvious. Although, as Chapter III
argues, the scientific axioms are more fundamental than the
ethical axioms, nevertheless in both cases we have propositions
which underlie more Specific claims and cannot themselves be

inferred from more fundamental claims.

152

Because of the problems of justification which are
built into the (strong) concept of intuitive knowledge, vari-
ous attempts have been made to find some other account of
such propositions, an account which would preserve their
epistemic status but eliminate the justification problem.

Some have attempted, for instance, to recast the a priori
scientific claims as rules of inference.1 Without entering
into the details of the debate, let me simply assume here

that Hempel is right in concluding that it is "simpler and
more helpful" to construe such laws as statements. Remember,
too, that for Sidgwick epistemology deals only indirectly with
principles of inference. Its prOper subject is the truth of
non-inferential principles. (See Chapter II of this disserta-
tion.)

Another alternative is simply to deny that these
fundamental principles constitute knowledge. A non-
foundationalist who makes such a denial (Regan, for instance)
can Still maintain that other propositions nevertheless count
as knowledge. When a foundationalist denies that the a priori
principles constitute knowledge the results are more extreme.
The history Of empiricism shows the extent Of the skepticism

which can result.

 

lCarl Hempel, As ects of Scientific Explanation (New
York: The Free Press, 1965), pp. 354-59.

 

153

A more useful approach to the tOpic occurs in Bertrand
Russell's Human Knowledge, Its Scope and Limits. His posi-
tion begins with the Observation that "'knowledge' . . . is
not a precise conception."l To refuse to call induction
knowledge is to reject a large portion of what the ordinary
person calls knowledge, and restricts the range of that word
very narrowly. At the very least it goes counter to ordinary
usage; in addition it lessens the possibility Of evaluating
and discriminating among various candidates for a priori
knowledge. Russell maintains that the restriction Of the
word knowledge to areas where we have complete certainty will
confine us to a complete solipsism of the present moment.2
There are, then, only two logically consistent positions:

either we must accept skeptical solipsism in its
most rigorous form, or we must admit that we know,
independently of experience, some principles or
principles by means of which it is pgssible to
infer events from other events . . .

Let us assume that this is an acceptable justification
for accepting a priori claims as intuitive knowledge in the
strong sense, in regard tO empirical scientific knowledge.

The resemblance of Sidgwick's ethical intuitionism

to Russell's treatment Of a priori empirical claims is clear.

As the first chapter Shows, it is Sidgwick's insistence on

 

1Bertrand Russell, Human Knowled e, Its SCOpe and
Limits (New York: Simon and Schuster, l 8), p. 96.

2Ibid., p. 178.

31bid., p. 179.

154

the cognitive status of moral claims which leads him to
intuitionism; and this parallels Russell's refusal to deny
the cognitive status of science. While Sidgwick never
reflects on the imprecision in the common notion of "knowl-
edge," he would surely agree on this point.

Russell, however, is no intuitionist--or at least,
he refuses to call himself that. The "principle or princi-
ples" which "we must admit that we know independently of
experience" he calls "necessary assumptions." His defense Of
them as playing a necessary role in his epistemology parallels
Sidgwick's defense of the necessity for second level axioms.
This parallel suggests that an overtly intuitionistic treat-
ment of such axioms is not absolutely demanded.

Sidgwick's axioms resemble Russell's a priori claims,
then, in that they underlie a whole field Of supposed knowl-
edge. On a foundationalist account, they must be knowledge
if any of the less general claims within the field are to be
knowledge. The "dependent validity" of third level prOposi-
tions demands a "superior validity" in second level. DO the
axioms resemble the a priori claims in lacking a decision
procedure?

Chapter II points out that Sidgwick's concept of self-
evidence allows him to call in supporting considerations; the
chapter argues that this concept is a coherent one. Therefore
the present question about kinds of intuition depends upon

the success of the considerations Offered in support of those

155

axioms. Chapter II raised some questions about those
considerations. The most interesting problems occurred in
his treatment of Prudence (as impartiality over time) and of
Rational Benevolence. These considerations involved mathe-
matical analogies. Chapter V looks at these supporting
analogies'again. What is clear is this: Sidgwick was seek-
ing intuitive foundations in the weak sense, foundations which
were Simply non-inferential. If in fact he accepts as axioms
principles which are non-inferential and for which there is
no decision procedure, it is because he could not find ade-
quate means Of showing the axioms to be self-evident.
Sidgwick is an epistemological intuitionist, then,
in regard to second level prOpositions, at least in a weak
sense. He is an intuitionist in a strong sense in his claim
that the only ultimate good is the pleasure of sentient
beings; he is an intuitionist in the strong sense about the
second level axioms of Obligation if he accepts axioms whose

self-evidence he cannot fully establish.

V. Conclusion

 

This chapter has attempted to sort out various stages
of the term ”intuitionism," noting in the process that some
of the variety of meanings attached to it result from the
frequent historical connection between deontology as an ethi-
cal position and intuitionism as an epistemological position.

When the two are separated and intuitionism is used in a

156

purely epistemological sense, it can then be asked whether
Sidgwick holds that propositions of the second or third levels
are intuited.

Sidgwick's treatment Of third level propositions
strongly resembles the Rawlsian technique of "reflective
equilibrium." After summarizing various problems in that
technique, both as used by Rawls and as apparently used by
Sidgwick, this chapter argues that upon closer examination
Sidgwick is not using reflective equilibrium and does not
treat third level propositions as intuitively known.

The more difficult and more significant question is
whether Sidgwick treats second level axioms as intuitively
known. Returning to the epistemological notion Of intuition-
ism, the paper distinguishes a weak and a strong interpreta-
tion, and then is able to recast the question. Does Sidgwick
claim that we are justified in believing the second level
propositions and at the same time that there are no accepted
procedures for resolving doubts about their truth? In the
case of Sidgwick's claims about ultimate good, the answer is
yes. In the case of the second level axioms of obligation,
the answer is less clear.

These axioms are claims to intuitive knowledge; but
whether in the strong or the weak sense depends upon the suc-
cess of the "arguments" supporting them. Sidgwick sought
intuitive knowledge in the weak sense, and may in fact be

calling upon axioms which can be intuited only in the strong

157

sense. In any case, he treats the axioms as others treat

the a priori claims underlying scientific knowledge. Sidg-
wick's ethical position is unaltered if the "intuited axioms"
are recast in the mold of Russell's "necessary assumptions."
This suggests the conclusion that Sidgwick's epistemological
intuitionism is demanded by his ethical cognitivism but not
embraced for its own sake. The following chapter continues
the examination Of Sidgwick's attempt tO found ethical knowl-
edge On propositions which are known intuitively only in the

weak sense.
5

CHAPTER V
PROBLEMS WITH SIDGWICK'S INTUITIONISM
I. Introduction

Chapter IV concluded that Sidgwick sought ethical
principles (second level propositions) which were intuitively
known in the weak sense. This kind Of intuition can be known
without supporting arguments; on the other hand, should
doubts arise about the truth of the proposition, such support
can be helpful in settling the doubts.

The conclusion that Sidgwick sought ethical knowledge
in this sense is consistent with the conclusion of Chapter II,
that Sidgwick's concept of self-evidence allowed for the use
of supporting considerations. It is also consistent with the
conclusion Of Chapter I, that Sidgwick's concept Of morality
demands that the most basic moral statements be either known
intuitively themselves, or be derivable from other statements
which are known intuitively. A statement which is intuitive
in the weak sense is one which is known directly itself, but
it is not impossible that the statement also be capable of
being known through inference from other statements.

Now the fact that Sidgwick offered supporting considera-

tions does not prove that he sought ethical foundations which

158

159

are intuitive in the weak sense only. Considerations can

also be offered for intuitions in the strong sense. Only I
know whether or not I am in pain; but if someone points out

to me that my face is contorted and I have been rubbing my
jaw, their comments may make me aware of pain that I had not
before noticed. In the case of intuition in the strong sense,
however, such considerations are never decisive. I could not
legitimately infer from their comments that I am in pain.

Such considerations in the case of intuition in the weak sense
can be decisive. I could legitimately infer from the comments
of others that I have two legs. Sidgwick's success in finding
axioms which are intuitive in the weak sense therefore depends
on the decisiveness of the considerations with which he sup-
ports them.

It might be objected that any decisive considerations
given to support intuition in the weak sense must ultimately
be appeals to intuition in the strong sense. However, Sidg-
wick's foundationalism only requires that the fundamental
premises of our knowledge be self—evident. However, Chapter
II argues that self-evidence is a partially psychological
term, such that a claim may be self-evident to some people
at some times, and known to others through argument. This
kind of foundationalism allows for epistemological variety
in its premises. A stricter foundationalism which character-

ized its premises as logically fundamental and underivable

160

might indeed have to resort to intuition in the strong sense;
Sidgwick's foundationalism does not.

In addition, when moral knowledge is treated from a
foundationalist perspective, it need not follow that funda-
mental moral principles are themselves underived. AS Chapter
I points out, they may be either self-evident or derived from
premises which are self-evident. SO even if it were true
that any foundationalism is ultimately driven to claim intui-
tive knowledge in the strong sense, it would not follow that
fundamental moral claims must be of that nature. And, most
importantly, treating these claims as intuitive knowledge in
the strong sense allows no way to solve the significant dis-
agreements among humans about such claims.

This chapter looks more closely at the kinds Of con-
siderations with which Sidgwick supports his second level
propositions. It assumes the conclusion Of Chapter IV, that
Sidgwick sought ethical foundations which are intuitive in
the weak sense. It examines the charges (some Of them his
own) that he was unable to find such foundations. Section II
considers a variety of such charges; Section III concentrates
on the problem which is by far the most significant, that of
the possible conflict between Rational Egoism and Rational
Benevolence. Section IV looks at some contemporary attempts
to provide the considerations which would eliminate this

possible conflict.

161

II. Problems in Finding Intuitive Foundations

Chapter IV pointed out Sidgwick's inability
(unacknowledged) to specify when a desire is a rational
desire. When he claims that the only ultimate good (the only
rationally desired ultimate end) is the pleasure of sentient
beings, he is making a claim which is intuitive in the strong
sense. The inability to make this a claim to intuitive
knowledge in the weak sense is the central problem in Sidg-
wick's ethics, for he claims that any rational method of
ethics must Specify an end. This, at least, is the view Of
Marcus Singer. If he is correct, then any adequate set Of
axioms of obligation, for Sidgwick, will have to specify an

end.
Rationally Chosen Ends

At the very outset of his major work, Sidgwick defines
a "Method of Ethics" as ”any rational procedure by which we
determine what individual human beings 'ought'--or what it is
'right' for them--to do."l

Marcus Singer, intrigued by this notion, follows its

use throughout the Methods of Ethics, and concludes "that

 

there is a fundamental incoherence in the key concept of the

methods of ethics . . ."2 He points out, first, Sidgwick's

 

lSidgwick, Methods, p. 1.

2Marcus Singer, "The Many Methods of Sidgwick's
Ethics," Monist, 58 (July, 1974):441.

162

frequent insistence that a "method of ethics" must Specify
an ultimate end Of moral action. This ultimate end serves
as an ultimate reason for what is done, and allows for moral
reasoning which shows the relation between a prescribed action
and the accomplishment of this end. This in turn gives
coherence to the method.

It is here that Singer finds a problem. He holds
that this position is inconsistent with Sidgwick's professed

neutrality among Methods of Ethics, for by these criteria

 

"intuitionism" is not a method at all. It has neither system
nor ultimate end, in spite of Sidgwick's occasional claims to
the contrary: "I have used the term 'intuitional' to denote
the view of ethics which regards as the practically ultimate
end Of moral actions their conformity to certain rules."1
Singer claims that the phrase "practically ultimate end" is

not clear; that it results from Sidgwick's preconception "that
there must be an ultimate end . . . if there is to be a method
of ethics."2

Even if Singer is correct in calling "practically
ultimate end" an unclear phrase, this is not sufficient to
establish that Sidgwick has abandoned his "professed neutral-

ity." It would be sufficient only if Sidgwick's position of

"rational end" is essential to his claim that Utilitarianism

 

1Sidgwick, Methods, p. 96.

2Singer, "Many Methods," p. 438.

163

has greater rationality than Common Sense Morality. Sidgwick
finds that Utilitarianism can supply principles of "ultimate
validity"--non-conflicting principles which are of use in
practical Situations. He finds that Common Sense Morality
cannot. This is an abandonment of neutrality only if the
requirements for "ultimate validity" are debatable. In any
case, it is not clear that Sidgwick's requirement that a
rational method of ethics specify an end plays any part in
his criticism of Common Sense Morality.

On the other hand, Singer has pointed to an important
problem in Sidgwick's thinking. If the phrase "practically
ultimate end" is not particularly clear, it may be because
the concept of "rationally chosen end" is not clear.

Singer's objection rests on the unclarity of "prac-
tically ultimate end." An "end" in this context supplies a
reason for action; and so problems about what is to count as
an end are related to problems about why one end is rational
and other ends are not.

Singer, in fact, attributes to Sidgwick the view that
the end which a method specifies need not be rationally chosen.
For this assumption he finds some support in the following
passages:

It is not necessary . . . to assume that the end
itself is determined or prescribed by reason: we

only require to assume . .1. that it is adopted
as ultimate and paramount.

 

1Sidgwick, Methods, p. 8.

164

The applicability of a method for determining

right conduct relatively to an ultimate end . . .

does not necessarily depend on the acceptance Of

the end as prescribed by reason; it only requires

that it should be in some way adOpted as ultimate

and paramount.
In both cases Sidgwick goes on to say that he has simplified
the task of his book by choosing to deal only with those ends
which may be considered reasonable. What is the significance
of those two remarks by Sidgwick? There are several possible
interpretations. At first glance, it is tempting to see
Sidgwick as oscillating--as he is accused of elsewhere--
between a descriptive and a normative position. On the one
hand, he acknowledges that peOple 99 act for non-rational
ends, but is unwilling to acknowledge that they Should so
act. There are, however, Other possible interpretations,
shedding more light on Sidgwick's basic stance.

The example he himself gives of a non-rational ulti-
mate end is Fame; and it is counted non-rational because upon
reflection we see that it is desirable not for itself but for
some other good, such as happiness. SO, many choices which
in practice function as ultimate, are upon reflection seen
as instrumental values only; and Sidgwick considers the way
we think rather than the ways we happen to act.

This puts some restriction on what is to count as a

rationally ultimate end. However, earlier in the book Sidg-

wick drops from consideration certain methods of ethics which

 

1Ibid., p. 77 n.

165

do prescribe ultimate ends. Methods based on the pursuit of
Excellence, of Perfection, and Of Virtue are all rejected.

SO, beyond the restriction to an ultimate end, we are left

in some doubt about what would distinguish a rationally chosen
and from others. It may be for this reason that Sidgwick is
reluctant to claim that the end of any prOperly so called
Method of Ethics is rationally chosen. This reluctance, how-
ever, coexists uneasily with Sidgwick's professed aim Of
seeking a rational foundation for morality.

I will assume here that a consistent and sympathetic
interpretation of what Sidgwick means by "method" includes a
rationally chosen ultimate end,1 as well as a rational decision
about which actions accomplish that end. These ultimate ends,
expressed by principles,2 must be mutually consistent, com-
plete, and related to one another.

It has been claimed that there is a different kind of
problem in Sidgwick's attempt to establish self-evident second
level propositions. This problem concerns the relationship

between the axioms of Obligation and Common Sense Morality.

 

1This interpretation is favored by many commentators.
William Frankena, "Sidgwick and the Dualism of Practical Rea-
son," and David Daiches Raphael, "Sidgwick on Intuitionism,"
Monist 58 (July 1974):452, 408.

2See Schneewind, Sidgwick, pp. 194-97, for a discussion
of the relation between principle and method.

 

166
"Unconscious Utilitarianism"

When Sidgwick summarizes what he holds to be the
relationship between Common Sense Morality and Utilitarianism,
he writes that ". . . the morality Of Common Sense may be
truly represented as at least unconsciously utilitarian."l
To call the Common Sense moralist an "unconscious utilitarian"
is provocative, almost metaphorical; it is less than completely
clear and has been interpreted in various ways. In the com-
memorative Monist volume, for instance, Gertrude Ezorsky
argues that Sidgwick's description Of what it would mean to
be an "unconscious utilitarian" is inadequate. She says the
term would be correctly explicated as follows:

An agent is an unconscious utilitarian if, and
only if:
(a) the view that one ought to maximize
utility is such that the agent doesn't
think himself to hold that view,
yet
(b) he acts as someone (with the same rela-
vant nonmoral beliefs) would act who
believed in maximizing utility.
Sidgwick, she claims, uses the following inadequate replacement
for (b):
(bb) A [Common Sense Moralist] in following
his moral rules tends to raise general
utility somewhat.
But (bb) does not entail (b); for instance, someone who chooses

to inflict a deserved punishment rather than an (apparently)

more utile undeserved punishment conforms to (bb) but not to

 

lSidgwick, Methods, p. 424.

167

(b). Since (bb) does not entail (b), then if (b) is part Of
the only adequate characterization Of ”unconscious utility,"
Sidgwick fails to achieve such a characterization.1
Now Ezorsky's explication of "unconscious utilitarian"

has a certain plausibility, and her claim that Sidgwick sup-
ports (bb) rather than (b) is correct. In his own words,

Utilitarianism is not concerned to prove the absolute

coincidence in results of Common Sense Morality and

Utilitarianism . . . nor to prove that existing moral

rules are mgge conducive to the general happiness than

any others; but only to point out in each case Some

manifest feliCific tendency which they possess.
And if the other premise of her argument sthnds--that hers is
the only adequate characterization Of the phrase "unconscious
utilitarian"--then her case is made. She can then go on--as
she does--to argue that in this sense the Common Sense Moralist
is not in fact an "unconscious utilitarian." This argument
resembles many which have been given over the last century,
whose point is that utilitarianism dictates actions which
common sense morality rejects.

But after all, the phrase "unconscious utilitarian“

is Sidgwick's; and Since it occurs in the same paragraph with

his disavowal of (b), there seems no good reason to claim that

any adequate explication of the phrase must include (b).

 

lGertrude Ezorsky, "Unconscious Utilitarianism," Monist
58 (July 1974):468-69.

2Sidgwick, Methods, p. 425.

168

What is at issue here is the complex relationship
which Sidgwick claims exists between Common Sense Morality
and Utilitarianism. He claims that Utilitarianism has a
rationality which is missing in Common Sense Morality. This
rationality accounts for the systematization and dependency
relationships that are explained in detail in Chapter IV.
Briefly, Sidgwick points out that Common Sense Morality is
consistent, and can claim general agreement, only when its
propositions are too general to apply to actual situations.
When its propositions are specific, they conflict with one
another and the resolution of the conflict is not the subject
of general agreement. These propositions are nevertheless
knowledge claims, and even have a "dependent validity." If
they are valid at all, that validity derives from their rela-
tionship to self-evident axioms which do meet the requirements
of consistency and applicability. Sidgwick believes that "the
apprehension, with more or less distinctness, Of these [second
level] abstract truths [is] the permanent basis Of the common
conviction that the fundamental precepts of morality are essen-
tially reasonable."l

Are arguments such as Ezorsky's effective against Sidg-
wick's claim, seen now in its complexity, that the common sense

moralist is an "unconscious utilitarian“? Certainly no single

counter-example--such as Ezorsky's instance Of common sense

 

1Ibid., p. 383.

169

morality choosing deserved punishment over apparently more
utile undeserved punishment--wi11 do the job. Sidgwick is
speaking of Common Sense Morality as a whole--as a complex
moral code. He claims that the system as a whole possesses
"manifest felicific tendency"--but he is not claiming that
each individual principle of that code is one a utilitarian
would choose to live by. Neither is he making the comparative
claim that our extant moral code is more felicific than any
conceivable alternative. In fact, it is the Utilitarian's
duty to rectify the moral code of his age. He expresses this
obligation in a nearly poetic passage near the end Of Methods.

[The Utilitarian] will contemplate [established

morality] with reverence and wonder, as a marvellous

product of nature, the result of long centuries of

growth . . . he will handle it with respectful

delicacy as a mechanism, constructed of the fluid

element of opinions and dispositions, . . . which

no politicians or philosophers could create . . .

Still, as this actual moral order is admittedly im-

perfect, it willlbe the Utilitarian's duty tO aid

in improving it.
And he is willing to accept, when required by his Utilitarian-
ism, moral judgments which run counter to received morality.
To cite one of the most problematic of such instances:

Thus on Utilitarian principles, it may be right to

do and privately recommend, under certain circum-

stances, what it would not be right to advocate

Openly . . . it may be conceivably right to do, . . .

if perfect secrecy can be reasonably expected, what

it wouldzbe wrong to recommend by private advice or
example.

 

1Ibid., pp. 475-76.
2Ibid., p. 489.

170

Sidgwick acknowledges the "paradoxicality" of these results.

There is no doubt that the moral consciousness of

a plain man broadly repudiates the general notion

of an esoteric morality . . . it would be commonly

agreed that an action which would be bad if done

Openly is not rendered good by secrecy.
He is fully granting, then, Ezorsky's claim that Common Sense
Morality sometimes requires acts which Utilitarianism forbids.
He grants this claim because it does nOt run counter to his
own position on the complex relationship between Common Sense
Morality and Utilitarianism.

But what would count against his conclusions? Theo-
retically, a number of things: if common sense morality could
be Shown to have the characteristics (roughly, exceptionless
applicability) which Sidgwick claims to be rationally neces-
sary, then there would be no need to look elsewhere for
rational principles; the "dependency" would not exist. This
is unlikely; the results of Sidgwick's analysis Of Common
Sense Morality, on the whole, are almost unarguable.

A second possibility would be the existence of an
alternative principle, or set of principles, which provided
greater systematization than Utilitarianism does; that is,
one which determined actual rightness and was exceptionless.2

If a competing principle or set Of principles provided less.

paradoxical results--that is, was more consistent with Common

 

1Ibid., p. 490.

2Given that such princip1e(s) satisfied the tests of
self-evidence as well as Utilitarianism does.

171

Sense Morality--such princip1e(s) would be of "superior
validity" to Utilitarianism and, again, Sidgwick's claimed
relationship would not hold. It is, then, not merely the
existence of a counter-example--a "paradoxical" conclusion

of utilitarian reasoning--but the existence Of counter-
examples together with an alternative set Of principles, which
would serve as evidence against Sidgwick's claimed "uncon-
scious utilitarianism."

There is still a third possibility, and one which,
unlike the first two, needs somewhat detailed consideration
here. If Sidgwick's version of Utilitarianism fails to meet
his own requirements for systematization, then, the dependency
and systematization relationship does not hold. One specific
possibility is a conflict between or among Sidgwick's second
level axioms; since he does not claim that any of the self-
evident axioms takes precedence over the others, the existence
of a conflict would destroy his claim to have provided excep-

tionless principles determining actual rightness.
Possible Conflicts Between Axioms

Could there be a conflict between his principles of
Justice and Benevolence? Claims that there can be such con-
flicts have long been a major part of arguments advanced
against utilitarianism. Two recent writers, A. R. Lacey and
James Newell, present interesting versions of this problem,

directed Specifically toward Sidgwick's version of

172

utilitarianism. Each author presents an imaginary Situation
in an effort to Show such a conflict; the two situations
differ, but they involve the same fundamental issue.1

In Newell's scenario, there is some good (understood
in Sidgwick's hedonistic sense) which occurs in indivisible
units, and moreover is in short supply; there may be, for
instance, three equally deserving candidates but only two
units Of the good to be given out. Newell argues that the
principle of Justice demands that no one receive the good,
since it is impossible for all who deserve it to receive it;
while the principle of Benevolence requires that we maximize
good, and therefore give the available goods to two people
rather than to none.2

In this fomm the Objection is fairly easy to meet;
for it is reasonable to claim that the principle of Justice
will be satisfied if these equally deserving candidates are
given an equal chance at receiving the good; and this can be
accomplished by some kind of random selection such as the

drawing Of lots.

 

1Newell is indebted to Lacey.

2James Newell, "The Philosophy of Henry Sidgwick,"
pp. 233, 304-6. Newell misdescribes the problem, for he
claims that "whether or not [the two principles] come into
conflict in actual practice, if they will issue in conflicting
duties when applied to a possible Situation, they are logically
incompatible principles." p. 304. This issue is discussed in
the following section of this chapter.

 

173

Lacey's scenario is more challenging. He deals not
with goods but with evils, namely pains; and he deals speci-
fically with a penal system. Suppose that the common good
is increased by the administration of some punishment, per-
haps because of its deterrent effect. Suppose, however, that
the increase in the common good would be outweighed by the
evil of punishment if everyone who deserved punishment received
it. The principle of Benevolence, then, demands that some, but
not all, Of the guilty be punished, while the principle of
Justice demands that all (supposing they are equally guilty)
be treated alike. Either all should be punished, or none
should be punished.1

But here, too, random selection would allow both
axioms to be satisfied. That this goes counter to our third
level convictions about proper penal institutions is not
directly relevant, for, as mentioned above, Sidgwick's posi-
tion allows for some direct contradiction between third level
convictions and second level axioms. In fact, his explicit
[acceptance of actions done in secret (where the same action
done publicly would be wrong) allows for the possibility of
a secret lottery in such a case. The fact that such a lottery
is almost by definition impossible to keep secret--for the
public would almost surely know that one was punished but not

another--reminds us of how unlikely it is that Lacey's

 

1A. R. Lacey, "Sidgwick's Ethical Maxims," Philosophy
34 (1959), pp. 217-28.

 

174

particular conflict between Justice and Benevolence would
ever take place in real life. Choices about penal systems
are choices between systems rather than between individual
actions--in this case, between a system Of uniform punishment,
arbitrary punishment, or uniform non-punishment. Arbitrary
punishment would cause less direct pain than uniform punish-
ment, for fewer people would receive punishment; but it would
cause more indirect pain, both in offering less effective
deterrence and in lessening the predictability Of public life.
It is unlikely, then, that a strictly utilitarian calculus
would call for public arbitrary punishment; but in any case,
punishment can be administered in such a way--impartial random
selection--as to satisfy both Benevolence and Justice.

But what of possible conflicts between Rational Bene-
volence and Rational Egoism? To what extent would the possi-
bility of such conflict destroy Sidgwick's claim to have
provided the systematization which is missing in Common Sense
Morality?

III. Possible Conflicts and the
DualiSm of Practical Reason

Of all that Sidgwick wrote, it is the Skeptical--
almost despairing--conclusion of Methods of Ethics which has
received the most attention. This despair arises from prob-
lems in reconciling Egoism and Utilitarianism.

Sidgwick has established, he thinks, that Intuitionism

(common sense morality) is a form of "unconscious utilitarianism,"

175

and asks whether there is a similar convergence between
Utilitarianism and the third "Method of Ethics," Egoism. The
“self-evident axioms" of earlier discussions are re-phrased
as principles--principles specifying ultimate ends--for the
two methods. The principle of Prudence (Pl)’ "One ought to
seek one's own good," becomes in the final chapter of Methods
the principle of Rational Egoism: "My own greatest happiness

is the rational ultimate end for me."1

This principle speci-
fies the ultimate end of the Egoistic Method. The principle
which Specifies the ultimate end of the Utilitarian Method
is Rational Benevolence.

AS stated in the final chapter Of Methods, this is,
"It is right and reasonable for me to do what I believe to be
ultimately conducive to universal Good or Happiness."2 Sidg-_
wick tells us that to accept the maxim of Rational Egoism while
rejecting that of Rational Benevolence is a reasonable posi-
tion; for the truth Of the one does not imply the truth Of the
other. Moreover, the coincidence of the dictates of the two
maxims is "at least incapable of empirical proof." The exis-
tence of a Moral Order which rewards those who abide by
Rational Benevolence and punishes those who do not, would

establish such a coincidence; for it "would suffice to make

it always everyone's interest to promote universal happiness

 

1Sidgwick, Methods, p. 497.

2Ibid., p. 507.

176

to the best of his knowledge."1

But the existence of such a
Moral Order--w ether as a personal God, or in some non-personal
form--cannot be proven.2
His resulting despair is stated most dramatically in
the First Edition:
The cosmos of Duty is thus really reduced to a
Chaos: and the prolonged effort Of the human
intellect to frame a perfect ideal of rational
conduct is seen to have been foredoomed to inevit-
able failure.
The above is a rough outline of Sidgwick's "Dualism
of Practical Reason." Sidgwick himself‘says very little about
this Dualism; he discusses it only in the final chapter of
Methods of Ethics (summarized above) and even there most of
his attention is given to showing that in ordinary life (with-
out the theological postulate) Egoism and Utilitarianism will
sometimes reach conflicting conclusions about the reasonable
thing to do. The nature of the conflict--its logical and
epistemological character--is not directly examined. Com-

mentators have interpreted the dualism differently. What

follows is an attempt to sort through various interpretations,

 

1Ibid., p. 506.

2Sidgwick, however, never doubted the possibility of
such knowledge. He Simply felt it had not yet been attained.

3Sidgwick, Methods of Ethiog, first ed., p. 473,
quoted in Schneewind, Sidgwick, p. 352. Sidgwick sometimes
calls this a conflict between Duty and Self-Interest; the
phrase is probably ill-chosen, since the issue involves a
tension between two duties: duty to self and duty to others.

 

177

in order to evaluate Sidgwick's success in finding self-evident

second level propositions,
Conflicting Interpretations

An historically important version of the dualism is
that of C. D. Broad. In Five Types of Ethical Theopy, he
writes:

It is surely quite plain that no such [theological]
postulate would free ethics from the theoretical
inconsistency . . . There are two principles which
are logically inconsistent . . . No God, however
powerful and however benevolent, can alter the fact
that those two principles are logically incompatible
and that therefore something which seemed self-
evident to Sidgwick must in fact have been false

. . . the only function of the postulate would be

to make it a matter of practical indifference whether
I acted in accordance with one or other Of t e two
principles, one of which must be false . . .

This conclusion is much stronger than the one Sidgwick reaches:

[Nlegation Of the connexion [of Benevolence and

Egoism] must force us to admit an ultimate and

fundamental contradiction in our apparent intui-

tions . . . it would seem to follow that the

apparently intuitive Operation of the Practical

Reason . . . is after all illusory. °
TO Sidgwick the two principles are possibly incompatible
rather than logically so; the possible incompatibility destroys
his hope of "completely rationalizing" ethics, but only the
actual incompatibility of the two principles would justify

denying their self-evidence. Actual incompatibility would,

 

1C. D. Broad, Five Types of Ethical Theory (New York:

Harcourt Brace, 1935), p. 253.

2Sidgwick, Methods, p. 508.

178

after all, mean that at least one of the principles was
false, and it would follow trivially that at least one of
the principles was not self-evident.

This, at least, is my interpretation of the passage
above; it is the only passage I know where Sidgwick addresses
this particular issue. (This interpretation is shared by
Schneewind.)l

Frankena takes an interesting position, midway between
that of Sidgwick himself and that of Broad. While he agrees

that the two principles are not logically incompatible, he

 

holds that even the possibility of their incompatibility

 

destroys their claim to self-evidence. 'These two questions--
of the logical relationship between the principles and of the
effect of that relationship on their self-evidence--deserve
attention.
The Logical Relationship Between
the Principles
Broad holds that the two principles are logically in-
compatiblel He chooses not to use any of Sidgwick's formula-
tions, and he paraphrases the two principles:
Egoistic Hedonism says: 'You ought to sacrifice
any amount Of happiness in others if you will
thereby increase your own total happiness . . .
more than you could by any other course of

action . . .' Universalistic Hedonism says: 'If
a certain sacrifice of your own happiness will

 

lSchneewind, Sidgyick, pp. 370-74.

179

so much increase that of others that the 39331

nett amount of happiness is increased, you ought

to make this sacrifice.‘
He adds Similar paraphrases concerning choices between actions
which lead to unhappiness. In what sense are these principles
"logically incompatible"? If two statements are logically
incompatible, then no matter what the state of the world it
is not possible for both to be true. It might be questioned
whether these principles are statements (and hence capable of
being true), but the slightest of rewordings will take care
of that problem; for instance,

Utilitarianism: If a certain sacrifice of your own

happiness will so much increase that Of others that the

total net amount of happiness is increased, making this

 

sacrifice is what you ought to do.

Egoism: If a sacrifice of ggy amount of happiness in

others will increase your own total happiness, this is

what you ought to do.
"X is what you ought to do" means "Doing X is right, and not
doing X is wrong." (Sidgwick himself goes back and forth
between formulations of these kinds. See Schneewind's com—
prehensive listing, pp. 295-297.) Now it is easy to see that
the two principles are incompatible--cannot both be true--
only in some states Of the world: those where doing the

action commanded by one principle requires a violation of the

 

1Broad, Five Types, pp. 240-41.

 

180

other. If there is a moral order in the universe--if
Sidgwick's "theological postulate" is true--then the two
principles are mutually consistent. If the theological postu-
late is false, then the two principles are inconsistent; how-

ever, even in this case the two principles are not logically

 

inconsistent. The inconsistency does not hold in all possible

worlds, but results fran a matter of contingent fact.
The Dualism and Self-Evidence

~ What about Frankena's claim that even the possible
conflict of the two principles destroys their claim to intui-
tive self-evidence? Where Broad concludes that it is impos-
sible (even granting the theological postulate) for both
principles to be true, Frankena says that it is impossible
(even granting the theological postulate) for both to be
self-evident. Frankena's conclusion is weaker than Broad's,
but stronger than Sidgwick's. (At least as I have interpreted
Sidgwick, he says that only the falsity of the theological
postulate would force us to deny the self-evidence of one of
the principles. Since we do not know whether that postulate
is true, we are left in suspense about the self-evidence Of
the two principles.)

Frankena says very little to support his position here;

he says simply that

it seems to me that two ethical principles cannot both
be regarded as self-evident if it is in principle

181

possible for them to come into conflict . . .
coincidence [in practice] might Obtain only
because Of a fortunate accident about the con-
stitution of our world, and not be true of other
possible worlds. [Emphasis added]

 

"But this is a hard question," he concludes; and it is. It
involves the distinction between the contingent and the neces-
sary. On one view--a completely deterministic one--there is
no difference between "contingent" states of affairs and
"necessary" ones. Everything is necessary, could not be
other than it is; there is only one ontologically possible
world. This view that the world is, ontologically, necessary
nevertheless allows room for epistemological contingency.
So-called "contingent" states Of affairs seem so because of
our ignorance Of the full, necessary interrelatedness between
what we are describing and the rest of the world.

Now most traditional theologians would call the exis-
tence of God, at least, necessary. Granting this assumption,
however, does not commit them to the view that every state
of affairs in the world (or in an after-world) is necessary.
For some traditional theologians, God has made the only pos-
sible world; these are committed to ontological determinism.
For other theologians, God has made one of many possible
worlds; these are not committed to ontological determinism.

Similarly, impersonal versions of the Moral Order postulate

 

1William Frankena, "Sidgwick and the Dualism Of Prac-
tical Reason," Monist, 58 (July 1974):457-58.

182

commit the holder neither to ontological determinism nor to
ontological contingency.

Would, then, an assumption that the existence Of the
Moral Order is necessary deprive Frankena's Objection of its
force? If the dictates of Egoism and Benevolence coincide,
this would not be contingent, but necessary. For that matter,
would assumptions like these create problems for the previous
contention that the two principles are not logically incom-
patible? (If they are incompatible when the world is fully
understood, such an argument could run, then they are incom-
patible by the very nature of things.) In either case, if
"contingent" retained a meaning, it would be epistemological
rather than ontological.

But this is just the point, for it is the epistemo-
logical status Of Egoism and Benevolence which is under dis-
cussion. With this reminder, problems about logical
incompatibility are obviated. With the understanding we now
have, and with any degree Of understanding which we are likely
to have, the principles of EgoiSm and Benevolence are at most
epistemologically incompatible, and that contingently so.

To return to the more puzzling question: Does our
uncertainty about the existence of the Moral Order destroy
the supposed self-evidence Of the two axioms? Self-evidence,
as discussed in Chapter II, is not limited by Sidgwick to
statements which are true by definition. It has to do, how-

ever, with epistemological independence, and is confined

183

roughly to those statements which can be known to be true as
soon as they are fully understood. Remember that Sidgwick
allows certain "demonstrations" of his self-evident axioms;
these involve, for instance, mathematical analogies in the
case Of Prudence, and "inference from" universalizability,

in the case of Justice. In the terms introduced in Chapter
IV, what Sidgwick means by "self-evidence" is intuitive knowl-
edge in the weak sense.

We cannot know that both Rational Egoism and Rational
Benevolence are true unless we know that there is a Moral
Order. The existence of such a Moral Order is not self-evident.
A statement cannot be intuitive in the weak sense unless it
is possible to know its truth without any supporting consid-
erations. That is not possible for the pair of axioms under
consideration. Therefore, it cannot be that both Of them are
self-evident, where "self-evident" is equivalent to "intui-
tive in the weak sense."

This conclusion is the same as Frankena's, although
it is reached by a somewhat different route. It is, as noted
before, a stronger conclusion than Sidgwick's own; it argues
that the dualism destroys the intuitive foundations Sidgwick
sought for his ethics. The conclusion is, however, milder
than Broad's, for it does not preclude the possibility of a
rational foundation for ethics, should some proof of the
theological postulate be found. Such rationalization, how-

ever, would take a somewhat different form than that imagined

184

by Sidgwick; it would either lack intuitive starting points
or have different ones than those he lists.

The most promising alternative rationalization is
suggested by attempts to Show that Benevolence has a claim
to self-evidence which is superior to that Of Egoism. The

next section of this chapter deals with that issue.

IV. Attacks on the Self-Evidence Of Egoism

If the principle of Egoism, that "my own greatest
happiness is the rational ultimate end for me," cannot be
intuitively known in the weak sense, then Sidgwick's Dualism
of Practical Reason no longer exists. There would no longer
be a conflict of two self-evident axioms.

Sidgwick's "proof" of utilitarianism would then be
much more plausible. His proof involves the claim that the
principle of Utilitarianism, Rational Benevolence, makes prac-
tical demands which do not conflict with the demands Of other
self-evident principles. He can then claim for Utilitarianism
the "superior validity" which allows Common Sense Morality to
have dependent validity.

The following sections of this chapter discuss several
attempts by contemporary writers to establish that egoism is
not a fully rational position, for such an attempt, if suc-
cessful, would salvage the intuitive foundations of Sidgwick's

ethical system.

185
Arguments from the Nature of Moral Language

A common position today is that the very word
"morality" contains within it a rejection of egoism; to call
an action moral and to call it egoistic is to contradict one-
self. The premises leading to this conclusion differ.
Frankena, for instance, argues quite directly: to be moral
is to be non-egoistic. He sees hints of such a position in
Sidgwick's tendency to use the term "ethical" when he is
referring to the principle of Egoism, and to reserve the term
"moral" for the non-egoistic principles.

Sidgwick, however, anticipated such an objection. He
answers that the general consciousness of mankind neverthe-
less considers self-interest reasonable, and his fundamental
search is for what is reasonable.

Kurt Baier, who also holds that the very meaning of
"morality" excludes egoism, argues from more complex premises.
Baier begins this way:

Consistent egoism makes everyone's private interest
the 'highest court of appeal.‘ But by 'the moral
point of view' we megg a point of view which is a
court of appeal for conflicts of interest. Hence
it cannot (logically) be identical with the point
of View of self-interest.

I claim, however, that egoism egg be that court Of

appeal. To the extent that the conflict exists within a

Single person, that is Obvious, for the conflict is resolved

 

1Kurt Baier, "The Moral Point of View," p. 470.

186

in favor of that person's own interest. If the conflict
exists between several people, egoism can even then meet
Baier's requirement of providing a solution which is mutually
acceptable, publicly reasonable. It can be mutually accept-
able that each person seek his or her own interest. This
contention will be developed more fully later.

But Baier's argument continues. One's own interest
is an aim, he says, but seeking it cannot be expressed as a
principle; yet "morality involves doing things on principle."1
What is it to act on principle? It is to act according to a
rule which one holds to be "binding on everyone alike quite
irrespective of what are the goals or purposes of the person
in question." This definition comes close to begging the
question, since it is in part Simply a direct rejection of
egoism. Egoists, after all, consider their own goals and pur-
poses as their ultimate end, and could not make a moral deci-
sion which disregarded these goals and purposes. Nevertheless,
this question of what is involved in acting on principle is
a crucial one. Objections to the rationality Of egoism
ordinarily invoke the necessary Objectivity, impersonality,
and universalizability of moral principles. Notice, for
instance, the following argument by Peter Singer:

If [one] claims to be aiming at what is morally

good, because of the universalizability inherent
in this claim, we can argue that he cannot,

 

1Ibid., p. 471.

187

consistently, give greater consideration to his

own interests, simply on the ground that they are

his own inferests, than he gives to the interests

of others.
The rest of this chapter explores the interconnections of
universalizability, rationality, and objectivity, and relates

them to egoism.
Reasonableness and Objectivity

To be reasonable means, of course, to be objectively
reasonable; that is, reasonableness involves meeting more
than private standards of rationality. Standards of ration-
ality are by definition Shared; an Objective reason is one
which can be shared. In moral discussions, an Objective
reason is a universalizable reason; that is, if it counts as
a reason for my doing an action, it would also count as a
reason for anyone's doing an action which is like it in
morally relevant ways. The point of many objections to the
rationality of egoism is that it cannot be expressed as a
universalizable reason for action. Sidgwick responds (al-
though not in exactly these terms) that egoism can be expressed
as a universalizable principle as long as it is expressed in
terms of subjective good rather than Objective good.

If the Egoist strictly confines himself to stating
his conviction that he ought to take his own happi-

ness or pleasure as his ultimate end, there seems
no Opening or line of reasoning to lead him to

 

1Peter Singer, "Reflective Equilibrium," p. 512.

188

Universalistic Hedonism . . . When, however, the

Egoist puts forward, implicitly or explicitly, the

proposition that his happiness or pleasure is Good,

not only for him but from the point of view Of the

Universe . . . it then becomes relevant to point ‘

out to him that his happiness cannot be a more

important part of Good, taken universally, than

the equal happiness of any other person.
The crucial distinction, in Sidgwick's mind, is between claim-
ing that something is good for oneself and claiming something
is good "from the point of view of the Universe." My para-
phrase just above this quote renders "good for oneself" as
"subjective good" and "good from the point of view of the
Universe" as "Objective good." This rendering is drawn from
an article by J. L. Mackie.

Mackie pinpoints the crucial importance Of this dis-
tinction in the defense of the rationality of egoism. (Sidg-
wick himself, here as in other discussions of the dualism, is
very brief.) AS Mackie sees it, "With his usual combination
of acuteness and honesty, Sidgwick has here got right some-
thing that many other moral thinkers have got wrong." What
is it that Sidgwick has got right? According to Mackie, it
is the distinction between "ethical objectivity in general"
(which Sidgwick treats as a given) and "the Objectivity Of

good." If the former can be maintained while the latter is

denied--and both Sidgwick and Mackie think it can--then the

 

lSidgwick, Methods, pp. 420-21.

189

reasonableness of egoism as an ethical principle can be
preserved.1

Mackie argues that statements about subjective good
are properly symbolized as two-place predicates, those about
Objective good as one-place predicates. He concludes that
the formal logical properties of the two concepts explain why
each Of the concepts is compatible with ethical objectivism;
that is, why one may accept the necessary Objectivity of ethi-
cal statements and nevertheless find egoism rational. If a
‘statement like "A's happiness is good" is universalized (and
ethical statements must be capable Of being universalized),l
the resulting statement is "The happiness of all persons is
good." "Good" here is a one-place predicate. Mackie provides
no symbolization, but it might look like this:

(1) "A's happiness is good" = Ga
(2) "Everyone's happiness is good" = (x)Gx

Sentence (2) is a universalization of sentence (1). To speak
of good as Objective, he says, is to treat it as a one-place
predicate. So, if the "good” in sentence (1) means "ultimate
good, a reason for acting," then anyone who holds sentence (1)
and also holds that ethical statements must be objective (uni-

versalizable) must hold that everyone's good is a reason for

acting. This last tenet is incompatible with egoism.

 

1J. L. Mackie, "Sidgwick's Pessimism," Philosophical
Quarterly 26 (October l976):3l9-20.

 

 

 

190

(3) "A's happiness is good for A" = Gaa

(4) "Each person's happiness is good for that
person" = (x)Gxx

Sentence (4) is a universalization of sentence (3). It pre-
serves the necessary Objectivity, in the sense of universal-
izability, of the statement, but it is not incompatible with
egoism.1

Hence Sidgwick's insistence on the distinction between
"good from the point Of view of the Universe" and "good for
oneself." Mackie's formal analysis helps us to see that
statements about the obligation to seek one's own good can
be universalized; and that this universalized formulation
expresses both an Objective sense of Obligation and a subjec-
tive sense of good, or of end. It helps us to see that egoism
can be defended against charges of being unreasonable which
identify reasonableness with the universalizability of ethi-
cal principles.

Mackie's analysis also helps us see the importance

of Sidgwick's distinction between objective good and

 

1My analysis here differs Slightly from Mackie's, who
casts statements about subjective good as "ought" statements,
but does not do the same for statements about Objective good.
His conclusion, then, depends on the claim that ought state-
ments must be symbolized with two-place predicates, while
”good" statements may be symbolized with either one-place or
two-place predicates. I cannot explain why he fails to con-
sider "ought" statements which refer to objective good.

191

subjective.l If the only reasonable way to Speak of "good"
is in the sense of "objective good," then egoism is unreason-
able, for egoism is incompatible with such claims. (See
Sentences (1) and (2) above.) If egoism is unreasonable,
then there no longer is a dualism in practical reason.

G. E. Moore claims that the notion of ”subjective
good" is unreasonable, and thus that the only reasonable way
to speak of "good" is in the sense of "Objective good." "The
only reason," he writes, "I can have for aiming at 'my own

good,' is that it is good absolutely that what I SO call

 

Should belong to me." This is the crux of Moore's rejection
of egoism, and it is part of his explicit attack on Sidgwick's
dualism. Moore argues that I can possess only something which
is good, "and not the fact that it is good." Things can be
possessed, but facts cannot; and, Moore goes on to say, sen-
tences like "The possession of this py_me is my good" are
meaningless.2

However, Moore fails to consider such closely related
sentences as "The possession of this is good for me." This
sentence removes the awkwardness of the one Moore chooses,

and furthermore expresses the idea which he rejects.

 

1From now on I will use Mackie's phrases rather than
Sidgwick's. Their parallel to one another is more Obvious
and the terms are less cumbersome.

2Moore, Principia Ethica, p. 98.

192

One further historical comment. C. D. Broad holds
the somewhat curious position that egoism is incompatible
with a purely teleological view of ethics. His argument is
as follows:

The consistent Egoistic Hedonist holds that pleasure

and nothing else is good, and that an equally pleasant

state is equally good no matter where it occurs. He

knows quite well that, in many cases, if he sacri-

ficed some of his own pleasure, others would gain far

more pleasure than he has lost. Yet he holds that

any such action would be wrong. Such a view would be

quite impossible if he held the teleological theory

that 'right' and 'conducive to intrinsically good

results' are mutually equivalent. It can be made

consistent only on the extreme deontological view that

such an action would be unfitting, and that its un-

fittingness suffices to make it wrong on the whole no

matEer how intrinsically good its consequences might

be.
The phrase "intrinsically good" is crucial here. Broad char-
acterizes teleology as the belief that a right action is one
which is conducive to intrinsically good results. He must
mean by "intrinsically good" what Mackie has meant by "objec-
tively good," for otherwise he would not find egoism and
teleology incompatible. He has assumed without noticing it
that the only reasonable way to Speak of ”good” results is
of "objectively good“ results.

Are there reasons other than Moore's for calling "sub-
jective good" an incoherent notion? These other reasons are

the ones to which this section Of the chapter is devoted.

 

lBroad, Five Types, p. 244.

193

Nagel and the Irrationality of
”Rational Egoism"
Thomas Nagel, in The Possibility Of Altruism,
advances an argument against the rationality Of subjective

concepts of good.1

This argument is important and interest-
ing in its own right; in addition, it has particular relevance
to the egoism which Sidgwick found rational. This relevance
stems from a position both men share about the rationality of
prudence (P2, the principle of impartiality over time). Sidg-
wick thinks that rationality demands impartiality toward all
the moments Of one's existence. But he also thinks that
rationality allows one to favor oneself over other peOple
(i.e., to take one's own good as an ultimate end and not take
the good of others as ultimate). Nagel agrees that ration-
ality demands impartiality toward all the moments Of one's
existence; but he does not think that rationality allows one
to favor oneself over other people. If Nagel begins with a
position so much like Sidgwick's but concludes with a rejec-
tion Of the rationality of egoism, it is interesting to con-
sider whether Nagel's conclusion is more faithful to their
common beginnings than is Sidgwick's.

The similarity in starting points is all the more

striking in the following passage. In it Sidgwick seems to

 

lThomas Nagel, The Possibility of Altruism (London:
Oxford University Press, 1970).

 

194

agree with Nagel that the same reasons which Show imprudence
to be irrational Show egoism to be irrational.
I do not see why the Egoistic principle should pass
unchallenged any more than the Universalistic. I
do not see why the axiom of Prudence should not be
questioned, when it conflicts with present inclina-
tion, on a ground Similar to that on which the Ego-
ists refuse to admit the axiom of Rational Benevolence.
If the Utilitarian has to answer the question, "Why
should I sacrifice my own happiness for the greater
happiness of another?" it must surely be admissible
to ask the Egoist, "Why should I sacrifice a present
pleasure for a greater one in the future?"
Sidgwick is not here speaking for himself; he is pointing to
the dilemma which the "extreme empirical school" must face.
Sidgwick is not an empiricist himself, and he has no trouble
rejecting imprudence as irrational. In doing so he con—
structs an argument which Nagel's later argument resembles.
Sidgwick's argument is quoted at length in Chapter II of this
dissertation; briefly, he says that the notion of "'Good on
the whole' of a single individual . . . is constructed by
. . . integration of the different 'goods' that succeed one
another" temporally. This is comparable to the notion of
"Universal Good" formed by "integration of the goods of all
individual human . . . lives." And yet, later, Sidgwick
finds egoism-~partiality toward one life--rational. He says

that

'I' am concerned with the quality of my existence
. . . in a sense, fundamentally important, in which

 

lSidgwick, Methods, p. 418.

195
I am not concerned with the quality of the existence
of other individuals.
He is unable to say more about this "fundamentally important
sense." If Nagel is right, this "fundamentally important
sense" cannot be a rational one.

Nagel finds Strict formal parallels in the rationality
of prudence and that of altruism. In his own words, "the
interpretation of that feature of reasons on which prudence
depends provides a model for the parallel enterprise in the
case of altruism."2 Nagel, then, focuses on the nature of
reasons, on what counts as a reason. This is a question
which Sidgwick never explicitly raises, and his failure to
do so leaves the central concept of "rationally chosen ulti-
mate end" unclear.

Nagel's argument, as I construe it, has the following
steps:

(1) Reasons, not desires, are necessary to explain
and justify action.

(2) Objective reasons motivate everyone who is aware
of them.

(3) Personal practical judgments have motivational

content .

 

1Ibid., p. 498.

2Nagel, Altruism, p. 27.

196

(4) Impersonal practical judgments must have the
same motivational content as the corresponding
personal practical judgment.

(5) Only reasons understood as objective allow
impersonal practical judgments to retain the
motivational content of corresponding personal
practical judgments.

Each of these steps is densely argued for, and the following
summary pays more attention to some of these arguments than
to others.

Step (1): Reasons, not desires, are necessary to
explain and justify action. One Of the most original sections
of The Possibility Of Altruism argues that desires are not
necessary to explain and motivate action. ("Desires" here
refers to presently existing feeling states.) Suppose, for
example, that I will visit Italy six weeks from now. At the
moment I have no desire to learn Italian; but the fact that
I will have such a desire in six weeks gives me a reason to
study Italian now.1 The book proceeds to examine the nature
Of reasons, as distinct from desires.

Step (2): Objective reasons motivate everyone who
is aware Of them. Nagel argues that both imprudence and ego-
ism fail to recognize an essential feature Of reasons, and

that in each case to deny this feature is to deny a related

 

lIbid., pp. 58-59.

197

empirical assumption. He argues that imprudence fails to
recognize that reasons are timeless, and that to deny their
timelessness is to deny that we are temporally extended
beings; egoism fails to recognize that reasons are Objective,
and this amounts to a denial that each of us is but one among
(at least possibly) many such beings.

Nagel stipulates a meaning of "Objective" which is
somewhat different from the uses of that term by authors con-
sidered earlier in this chapter. "Objective" is, of course,
Opposed to "subjective." A subjective reason can motivate us
only if we know our place in what it describes. An objective
reason motivates us whether or not we know our place. Nagel
expresses the difference formally:

Every reason can be expressed by a predicate R,

such that for all persons p and events A, if R is

true of A, then p has prima facie reason to promote

A . . . Formally, a subjective reason is one whose

defining predicate R contains a free occurrence of

the variable p (. . . free only within R; it will

be bound by the universal quanitification pver

persons which governs the entire formula).
He then gives the following examples of reasons someone--G. E.
Moore, say--might give for removing himself from the path of
an oncoming truck.

Objective reasons (intuitively formulated):

(a) the act will prolong G. E. Moore's life.

(c) the act will prolong someone's life.

The same Objective reasons (formally stated)

 

1Ibid., p. 90.

198

(a) (p,A) (If A will prolong Moore's life, then p
has reason to promote A.)
(c) (p,A) [If (3g) (A will prolong q's life), then
p has reason to promote A.]
These objective reasons motivate whether or not one is G. E.
Moore. Whoever I am, I have reason to do what I can to
remove Moore from the path Of the truck. This is not true
of subjective reasons, which do not motivate unless I know
that I am Moore.
Subjective reasons (intuitively formulated):
(b) the act will prolong his life.
The same subjective reason (formally stated)
(b) (p,A) (If A will prolong p's life, then p has
reason to promote A.)l
However, this formulation seems to make all prudential reaSons
subjective. Prudential reasons are those which motivate one
to take care of oneself. If these are subjective, then Nagel
must abandon his crucial claim that any reason, ggg reason,
is objective.2 Nagel deals with this in terms of "a curious
value: the Objective value of self-reliance in matters of
basic survival and self-preservation." He gives the follow-

ing intuitive and formal versions of such prudential reasons:

 

1Ibid., p. 91.

2See, for instance, the criticism made by Benson,
Philosophical Quarterly 22 (January 1972):82-83.

199

(d) the act is one which will prolong the life of
its agent.
(d) (p,A) (If A is an act which will prolong the
life Of its agent, then p has reason to promote
A.)1
In this characterization of "objective," Nagel believes
he has a tool which will Show that "value must be conceived as
Objective." This would be an argument for altruism, "Since
it will mean that the ends are common rational objects of
pursuit for everyone."2 But how does he support his position?
SO far I have simply described the difference between subjec-
tive and objective reasons. There remains the major task of
establishing that all reasons (as distinguished from desires)
are Objective--all reasons motivate everyone who is aware of
them to bring about the state of affairs for which they are
reasons.
(3) Personal practical judgments have motivational
content.
(4) Impersonal practical judgments must have the
same motivational content as the corresponding
personal practical judgments.
His argument for this conclusion depends upon another

distinction, that between personal and impersonal practical

 

1Ibid., p. 92.

2Ibid., p. 97.

200

judgments. Since we are not solipsists, since we see others
as persons like ourselves, we must be able to "say of other
persons anything which one can say of oneself, and in the
same sense." Motivational content is part of any personal
practical judgment; therefore, it must be a part of any im-
personal practical judgment. TO make a personal practical
judgment ("I have reason to do A") is to commit oneself to a
corresponding impersonal judgment ("Smith has reason to do
A”) plus some statement identifying myself as Smith. Since
the first judgment is sufficient to justify and explain my
doing A (this is what Nagel means by "motivational content")
so the second is sufficient to justify and explain action
bringing about Smith's doing A.1
His final premise is this:
(5) Only reasons understood as Objective allow
impersonal practical judgments to retain the
motivational content of corresponding per-
sonal practical judgments.
If all five Of these premises are true, Nagel can conclude
that all true reasons, ggg reasons, are Objective. But ego-
istic reasons are not objective, for egoistic reasons motivate

us only if they refer to us. Therefore, egoistic "reasons"

are not fully reasons.

 

1ibid., p. 101.

201

Nagel's position is not without problems. Most
attacks on it concern his claim that motivational content is
part of practical judgments. "The crux Of the matter is the
legitimate content of the claim that one should be able to
say 'I have a reason' and"He has a reason,‘ using 'reason'

in the same sense."1

Both Watt and Darwall2 have misgivings
about that "legitimate content." The following objection is
my own, but heavily indebted to both men. The problem seems
to be this: The personal judgment has motivational content
in that it explains and justifies my action; the impersonal
judgment has motivational content in that it explains and
justifies Smith's action. The word "reason" is therefore
used in the same sense. If the two judgments in question were
re-phrased:

1. There is a reason for me to do A,

2. There is a reason for Smith to do A,
it could plausibly be claimed that both have the same moti-
vational content; whoever makes either judgment explains and
justifies those actions of his which bring about Smith's

doing A. But these re-phrasings beg the question, for their

very wording assumes that reasons are Objective.

 

1A. J. Watt, review of The Possibility Of Altruism,
by Thomas Nagel, in Australasian Journal Of Philosophy, 48
(December l970):408.

 

2Stephen Darwall, "Nagel' s Argument for Altruism,"
Philosophical Studies, 25 (1974): 125- 130.

202

It seems, then, that egoism has not yet been proven
irrational. Nagel's attempt at such a proof is significant

because it deals with the problem which, as Sidgwick points

out, is central: the distinction between subjective and Objec-

tive reasons for action. If only the latter are truly
rational, then egoism has no rational basis, and the dualism
does not exist. Without the dualism, Sidgwick's effort to
find intuitive (in the weak sense) bases for ethics might
have succeeded.

Not only does Nagel attempt to establish what Sidg-
wick would like to have established; the premises from which
Nagel argues are of the kind which Sidgwick sought. These.
premises themselves claim to be intuitive in the weak sense.
Not that Nagel uses that particular terminology; but he does
claim that, when we fully understand the nature of reasons,
that very understanding will lead us to reject egoistic "rea-
sons" as irrational. Just as Sidgwick hoped that the con-
siderations he advanced in support of the self-evident axioms
were decisive, so Nagel hopes that the considerations he
advances in favor of the Objectivity of reasons are decisive.
The hOpe that supporting considerations will be decisive is
not available to one who depends upon foundations which are
intuitive in the strong sense.

The final pages of Chapter I argue that Sidgwick's
concept of moral knowledge demanded that the most basic moral

statements be known intuitively themselves, 93 that they be

203

derivable from some other self-evident, non-empirical
statements. Nagel attempts to supply such other premises.
Nagel, I argue in Chapter V, fails. But anyone who rescues
Nagel rescues Sidgwick.

In the meantime we must repeat Mackie's accolade:
"With his usual combination Of acuteness and honesty, Sidg-
wick has here got right something that many other moral
thinkers have got wrong." Until someone can establish that
only Objective concepts of good, of reasons for acting, are
rational, then Sidgwick is correct in claiming a dualism in

practical reason.

V. Conclusion

 

This chapter examines Sidgwick's effort to establish
second level propositions as intuitions in the weak sense.
It argues that the concept of a rationally chosen end is
integral to his concept of a method of ethics; and it recalls
that he calls one set of apparently self-evident ethical
principles superior in validity to another if the former is
exceptionless and can determine actual rightness, to a greater
extent than the latter. Most of the chapter deals with the
failure of Sidgwick's own "self-evident" principles to be
exceptionless. It is claimed that Benevolence and Justice
do not conflict, but that the possibility of conflict between
Egoism and Benevolence destroys their claims to self-evidence.

The two principles are not logically incompatible, but it is

204

possible for both to be true only if a theological postulate
is true. If they actually conflict, Obviously at least one
is false and therefore not self-evident; if they do not con-
flict (because the theological postulate is true), it still
cannot be claimed that each is self—evident since the truth
of at least one depends upon the truth of the theological
postulate. An intuitive statement must be capable Of being
known without supporting considerations.

The chapter then discusses a number of current
attempts to Show that egoism is irrational. The success of
such an attempt would salvage Sidgwick's claim to have found
at least one self-evident principle (Benevolence). But Egoism
survives the challenges raised against it, as Sidgwick had
claimed it would. His own brief remarks about egoism antici-
pate the more technical current discussions of objective and
subjective good and point out clearly that the essential ques-
tion is the legitimacy of the cOncept Of subjective good.

The fact that Sidgwick's comments Often presage cur-
rent discussion--and that these discussions shed light upon
his own position--allows this final conclusion: Sidgwick's
"intuitionism” is far more in the spirit of the twentieth
century than of the eighteenth. He holds, not that he has
found the self-evident bases of morality, but that claims to
have found such bases are properly subject to rational evalu-

ation.

CHAPTER VI

CONCLUSIONS

Sidgwick characterizes an adequate ethical theory as
one whose second level axioms are themselves self-evident
and which explain and justify the claims of third level
propositions to be knowledge. His search for such ethical
foundations is thorough; his evaluation of his own results
is honest and despairing.

Does the failure of such a careful and candid theo-
ritician show that all similar searches will end in failure?
If so, the efforts Of many twentieth-century ethicists are
doomed, for many are attempting (in far less detail, usually)
what Sidgwick attempted before them.

The true value of studying Sidgwick does not lie in
realizing the extent of his failure. Instead, it lies in
recognizing and making use of his successes, and in treating
his failure as the central problem which any foundationalist
ethics must solve.

His successes include a number of useful distinctions
among kinds of intuition, a modest and reasonable theory of
self-evidence, and an attitude toward common sense morality
which grants it heuristic value without calling it authori-

tative.

205

206

He was unable to find ethical foundations which
satisfied his own requirements of rationality. But his
attempt to do SO isolated a singularly important issue:

What counts as a reasonable end for action? Is an egoistic
concept of gOOd--"my good in preference to the good of
others"--such an end?

Although his work, for a careful reader, draws atten-
tion to this question, Sidgwick says very little which is of
help in answering it. He simply says that he does not know
how "subjective good" could fairly be called unreasonable.
Contemporary work on the nature of reasons is far more exten-
sive than anything Sidgwick considered. If it should succeed
in delineating a concept of "reason" which disqualified ego-
ism as a rational principle, this work would complete a
project which Sidgwick began.

Methods of Ethics does not Show the futility Of

 

foundationalist ethical theory. Instead, it points out the

major challenge which such theories must meet.

BIBLIOGRAPHY

BIBLIOGRAPHY

Alston, William P. "Self-Warrant: A Neglected Form of
Privileged Access." American PhilOSOphical Quarterly,
13 (October l976):257-272.

Baier, Kurt. TheMgral Point of View. Ithaca: Cornell
University Press, 1958.

 

Benson, John. Review of The Possibility of Altruism, by
Thomas Nagel. Philosophical Quarterly, 22 (January
1972):82-3.

Blanshard, Brand. "Sidgwick the Man." Monist, 58 (July
l974):349-70.

Bradley, F. H. "Mr. Sidgwick on Ethical Studies." Mind,

Broad, C. D. Five Types of Ethical Theory. New York: Har-
court, Brace, 1935.

Darwall, Stephen L. "Nagel's Argument for Altruism." Philo-
sgphical Studies, 25 (1974):125-30.

 

. "Pleasure as Ultimate Good in Sidgwick's Ethics."
Monist, 58 (July 1974):475-89.

Ezorsky, Gertrude. "Unconscious Utilitarianism." Monist, 58
(July 1974):468-74.

Frankena, William K. "Sidgwick and the Dualism of Practical
Reason." Monist, 58 (July 1974):449-67.

Gert, Bernard. Review of The Possibility of Altruism, by
Thomas Nagel. Journal Of_Philosophy, 69 (June 15,
1972):340-44.

 

Hare, R. M. "Rawls' Theory of Justice." In Reading Rawls,
pp. 82-98. Edited by Norman Daniels. New York:
Basic Books, 1976.

Havard, W. C. Henry Sidgwick and Later Utilitarian Political
Philosophy. Gainesville: University of Florida Press,

 

207

208
Hempel, Carl. Aspects of Scientific Explanation. New York:
The Free Press, 1965.

Hume, David. A Treatise of Human Nature. Edited by L. A.
Selby-Bigge. Oxford: Clarendon Press, 1967.

Lacey, A. R. "Sidgwick's Ethical Maxims." Philosophy, 34
(l959):217-28.

 

Lyons, David. "Nature and Soundness of the Contract and
Coherence Arguments." In Reading Rawls, pp. 145-59.
Edited by Norman Daniels. New York: Basic Books,
1976.

Mackie, J. L. "Sidgwick's Pessimism." Philosophical Quar-
terly, 26 (October l976):317-27.

Moore, George Edward. Principia Ethica. Cambridge: Cam-
bridge University Press, Cambridge, 1903.

Nagel, Thomas. The Possibility Of Altruism. London: Oxford
University Press, 1970.

Newell, James David. "The Philosophy of Henry Sidgwick."
Ph.D. dissertation, University of Maryland, 1975.

Raphael, David Daiches. Moral Jud ent. Westport, Connecti-
cut: Greenwood Press, 1978.

. "Sidgwick on Intuitionism." Monist, 58 (July 1974):
405-19.

Rawls, John. A Theory of Justice. Oxford: Clarendon Press,
1971.

Reba, Marilyn. "A Second Look at Nagel's Argument for Altru-
ism." Philosophical Studies, 25 (1974):429-34.

Regan, Tom. "McCloskey On Rational Ends: The Dilemma Of
Intuitionism.” Journal of Value Inquiry, 9 (Spring
1975):48-58.

Reid, Thomas. Essays on the Intellectual Powers of Man.
Cambridge, Massachusetts: TEe M.I.T. Press,*l969.

 

Rorty, Richard. "Intuition." In Encyclopedia of Philosoph ,
Volume 4, pp. 204-12. Edited by Paul Edwards. New
York: Macmillan CO. and The Free Press, 1967.

Russell, Bertrand. Human KnoyledgeLIts Scope and Limits.
New York: Simon and Schuster, 1938.

209

Schneewind, J. B. "Henry Sidgwick." In Encyglopedia Of

 

Philosophy, Volume 7, pp. 434-36. Edited by Paul
Edwards. New York: Macmillan Co. and The Free Press,
1967.

. "Sidgwick and the Cambridge Moralists." Monist,
58 (July 1974):37l-404.

. Sidgwick's Ethics and Victorian Moral Philosophy.
Oxford: Clarendon Press, 1977.

Schultz, Robert C. "Sidgwick on Proof in Ethics." Philo-

sophical Archives, September 1975, pp. 336-51.

Sidgwick, Henry. "Is the Distinction Between 'Is' and 'Ought'

Singer,

Ultimate and Irreducible?" Proceedings Of the
Aristotelian Soceity, 2 (1892-94):88-92; reprint ed.,
New York: Johnson Reprint Corp., 1963.

. "The Establishment of Ethical First Principles."
Mind, 4 (January 1879):106-11.

. "Hedonism and Ultimate Good." Mind, 0.5. 2 (Janu-
ary 1877):27-38.

. "Incoherence of Empirical Philosophy." Mind, 0.3.
7 (October 1882):372-90.

. "Criteria of Truth and Error." Mind, n.s. 9
(January 1900):

. Lectures on the Ethics of T. H. Green, Mr. Herbert
SpencerL and J. Martineau. New York: Macmillan Com-

pany, 1902; reprint ed., New York: Kraus Reprint,
1968.

. Lectures on the PhilOSOphy Of Kant. London: Mac-
millan and Company, 1905.

. Methods of Ethics. 7th edition. London: Mac-
millan and Company, Ltd., 1907; reprint ed., New York:
Dover Publications, 1966.

. Miscellaneous Essays and Addresses. New York:
Macmillan and Company, 1904.

. Philosophy, Its Scope and Relations. New York:
Macmillan Company, 1902.

Marcus. "The Many Methods of Sidgwick's Ethics."
Monist, 58 (July 1974):420-48.

210

Singer, Peter. ”Sidgwick and Reflective Equilibrium." Mon-
ist, 58 (July 1974):490-517.

Taylor, Frank Simmons. "The Right and the Good in the Ethics
of Henry Sidgwick." Ph.D. dissertation, Boston Uni-
versity Graduate School, 1969.

Watt, A. J. Review of The Possibility Of Altruism, by Thomas
Nagel. Australasian Journal of Philosophy, 48 (Decem-

ber l9705:405-09.

i ll illl'. '1...

"7111111111111lilies