. 2
( 8)


immediate awareness of good. For another, it has to be impressive if his
theory can provide the means to diagnose the mistakes of earlier philoso-
phers and retrieve and properly place their insights. In his not very schol-
arly but always interesting and provocative history of ethical theory,
Moore attempts just this kind of diagnosis. At the heart of that history, we
find, not indefinability, but nonnaturalness.
Moore begins his history by dividing ethical theorists into two camps,
the naturalists proper “ the empiricists “ and the metaphysicians.24 What
divides these camps are the different ways they have of inadequately treat-
ing good™s nonnaturalness. At one point, Moore suggests that the natu-
ralistic fallacy is an instance of a more general problem, the mistreatment
of all nonnatural properties.25 The empiricists crudely deny that there
are any such properties, while the metaphysicians, being subtler, do at
least recognize the category of nonnaturals, even if they then try to go
back on it.26 The mistreatment of good always starts with the same mis-
take. The vast majority of propositions we entertain have existents as both
subject and predicate: “Ethical truths are immediately felt not to conform
to this type and the naturalistic fallacy arises from the attempt to make
out that, in some roundabout way, they do conform to it.” Philosophers
do this by supposing that a nonnatural property “necessarily exists together
with anything with which it does exist.”27 Moore™s emphasis on “neces-
sarily” is not meant to suggest that there is a mistake in supposing neces-
sity to be involved in the being of nonnaturals. Two plus two is necessar-
ily four and friendship is necessarily good. The mistake lies in supposing
that the necessity has anything to do with existence. If our understanding
of nonnatural properties gets tied in any way to existence, we will in-
evitably fall prey to understanding them in terms of natural properties,
the ones that really exist. We will then come to think that what is true
about nonnatural properties depends on what exists.
The only way the empiricists have of accounting for the necessity of
propositions concerning nonnaturals is in terms of the things that have
actually existed. Thus they cannot make room for counterfactual propo-
sitions “ they are even forced to say that mathematical propositions would
not be true without the existence of just these things that have existed!28
So saying that the metaphysicians do a slightly better job than the em-

24 25 26
Principia, p. 124. Ibid., p. 125. Ibid., pp. 110“12.
27 28
Ibid., p. 124. Ibid.
28 g. e. moore™s ethical theory

piricists of dealing with nonnatural properties is damning with very faint
praise indeed. Refusing to accept nonnatural properties for what they
are, properties that do not exist in time, metaphysicians attempt to ex-
plain away their special character by defining them in terms of a super-
sensible reality.29 Even if officially it is “more” real than the world of natu-
ral objects, when the smoke clears and the mirrors are removed, this re-
ality is revealed to be a pale reflection of the natural one “ its being
sensible trumps its being “super.” When it comes to a mathematical
proposition, for instance, metaphysical philosophers “have no better
account of its meaning to give than either, with Leibniz, that God™s mind
is in a certain state, or, with Kant, that your mind is in a certain state, or
finally with Mr Bradley, that something is in a certain state.”30 In ethics,
instead of simply saying, as a naturalist might, that to be good is to be
pleasing to some person or another, the metaphysicians say that it is to
be pleasing to God, or to a rational will. This greater subtlety gives them a
subtler way of imperiling what is as plain as 2 + 2 = 4. They halfway man-
age to convince themselves that they must wait upon such superfluities as
whether God exists or whether history embodies a rational principle be-
fore they can conclude that love is better than hatred.
What we think, including what we think we think, affects how we feel.
So the errors of those who think incorrectly about value are of the great-
est moment. Consistent hedonists do what Mill could not quite get him-
self to do, recommend piggishness if that is what brings pleasure. Meta-
physicians, thinking that the truly valuable things can only be found in a
world which, even if it mirrors this one, somehow lies beyond it, either
become fatalists about this world or suffer some form of religious or quasi-
religious dissatisfaction that leaves them yearning to escape its too solid
flesh.31 Moore™s final message is one naturalists already profess but do not
understand. So he directs it instead to the metaphysicians, who he thinks,
however misguided, to be superior: Although it is not the best world pos-
sible, this is the world that justifies existence.

The Argument for Indefinability
Let us turn now to the argument by which Moore purports to prove that
good is indefinable, the OQA, along with the Private Language Argument
the most famous and influential “argument” of the century. We begin by
presenting at length the passage in which Moore makes his argument:
(1) The hypothesis that disagreement about the meaning of good is disagree-
ment with regard to the correct analysis of a given whole, may be most plainly

29 30 Ibid., p. 125.
Ibid., pp. 111“12.
Ibid., p. 205. We discuss Moore™s attempt to deal with this kind of disappointment in the
book™s last two chapters.
simplicity, indefinability, nonnaturalness 29

seen to be incorrect by consideration of the fact that, whatever definition be of-
fered, it may be always asked, with significance, of the complex so defined,
whether it is itself good. To take, for instance, one of the more plausible, because
one of the more complicated, of such proposed definitions, it may easily be
thought, at first sight, that to be good may mean to be what we desire to desire.
Thus if we apply this definition to a particular instance and say ˜When we think
that A is good, we are thinking that A is one of the things which we desire to de-
sire,™ our proposition may seem quite plausible. But, if we carry the investigation
further, and ask ourselves, ˜Is it good to desire to desire A?™ it is apparent, on a lit-
tle reflection, that this question is itself as intelligible, as the original question, ˜Is
A good?™ “ that we are, in fact, now asking for exactly the same information about
the desire to desire A, for which we formerly asked with regard to A itself. But it
is also apparent that the meaning of this second question cannot be correctly an-
alyzed into ˜Is the desire to desire A one of the things which we desire to desire?™:
we have not before our minds anything so complicated as the question ˜Do we
desire to desire to desire to desire A?™ Moreover any one can easily convince him-
self by inspection that the predicate of this proposition “ ˜good™ “ is positively dif-
ferent from the notion of ˜desiring to desire™ which enters into its subject: ˜That
we should desire to desire A is good™ is not merely equivalent to ˜That A should
be good is good.™ It may indeed be true that what we desire to desire is always also
good; perhaps, even the converse may be true: but it is very doubtful whether this
is the case, and the mere fact that we understand very well what is meant by doubt-
ing it, shews clearly that we have two different notions before our minds.
(2) And the same consideration is sufficient to dismiss the hypothesis that
˜good™ has no meaning whatsoever. It is very natural to make the mistake of sup-
posing that what is universally true is of such a nature that its negation would be
self-contradictory: the importance which has been assigned to analytic proposi-
tions in the history of philosophy shews how easy such a mistake is. And thus it is
very easy to conclude that what seems to be a universal ethical proposition is in
fact an identical proposition; that, if, for example, whatever is called ˜good™ seems
to be pleasant, the proposition ˜Pleasure is the good™ does not assert a connec-
tion between two different notions, but involves only one, that of pleasure, which
is easily recognised as a distinct entity. But whoever will attentively consider with
himself what is actually before his mind when he asks the question ˜Is pleasure
(or whatever it may be) after all good?™ can easily satisfy himself that he is not
merely wondering whether pleasure is pleasant.32
It is breaking no new ground to point out that Moore™s presentation of
the OQA is quite muddled. In fact, it is shocking how slapdash he is with
something he considers so important. The standard and most plausible
reading of the passage breaks it into three parts, the first being just the
first sentence, the second stopping at the sentence beginning with “More-
over,” and the third going perhaps all the way to the end of the second
paragraph, which we have not quoted in full. It then joins the first and
third parts together to form the main argument and treats the second
part as containing either a subsidiary argument or a non sequitor.
This reading ignores Moore™s own verbal cues, which suggest that he
considers the beginning of the third part to be not an argument, but

Principia, pp. 15“16.
30 g. e. moore™s ethical theory

rather a phenomenological appeal in support of the argument immedi-
ately preceding it. But the concern with differences in the significance of
questions that is expressed in the first and third parts gives them a great
deal more in common than either has with the second part, concerned
as it is with differences in the complexity of questions. Further, no per-
fectly general point can be plausibly assayed about differences in the com-
plexity of second-order questions. Assume, for instance, that good is as-
serted to be identical with a simple like pleasure. Even if we are confident
that the question whether A™s goodness is good has a different meaning
than the question whether A™s being full of pleasure is good, (we may be
too puzzled by the first of these questions to have any opinion about what
it means), there is no difference in the complexity of these questions, just
as there is no difference in the complexity of the parallel first-order ques-
tions concerning just A. Since, as we shall see, Moore expresses confusion
elsewhere about the difference between arguments and phenomenolog-
ical appeals, it does not speak against the standard view that it has him
not completely understanding his own strategy. It is also to be noted that
he never publicly opposed what quickly became the standard reading of
the argument.33
The argument can be pithily stated. The answer to a question whether
a property purported to be the defining one for ˜good™ actually is good
is always “significant,” or nontautologous. Since the answer to a question
whether something asserted to be good is good is tautologous, good and
the property purported to be its definition must be different. In the third
section, Moore provides the argument with its explication. Two different
things can be meant by the claim that a property is good: It can be meant
1) that the property is identical with good or 2) that the property has the
further property of being good. This is to say that the ascription of good-
ness to something can be either analytic or synthetic. Those who try to
define good take a significant, synthetic proposition and treat it as ana-
lytic. The lack of equivalence between a statement that is really analytic
and one that is mistakenly treated as such is shown by the fact that while
we cannot doubt the former, we can doubt the latter. Moore might have
made this point more clearly by contrasting the significance of the fol-
lowing pair of statements: “A, which is good, is good” and “A, which is
what we desire to desire, is good.”
In the second paragraph, where, officially, he considers and rejects the
alternative “that ˜good™ has no meaning whatsoever,” that is, that there is
no such thing as good, Moore continues the point and draws out the fur-
ther pernicious consequences that stem from the attempt to define ˜good™.
When he does so, he ensnares those who try to define not just ˜good™, but

Richard A. Fumerton, Reason and Morality (Ithaca and London: Cornell University Press,
1990), p. 70.
simplicity, indefinability, nonnaturalness 31

any concept or property, in that most famous of philosophical conun-
drums, the paradox of analysis. Moore™s conditions of analytic adequacy
require all analyses to be either trivial or false. An analysis states that an
identity obtains between the object denoted by the analysandum and the
object denoted by the analysans. If the identity does obtain then by the
terms of Moore™s argument, the analysis is trivial, stating no more than a
tautology. But if the analysis is nontrivial, “significant,” there is no iden-
tity between the two and the analysis must be false. In Principia, Moore
evinces no awareness that the sword he wields against those who define
˜good™ is double-edged. He fails to consider that the argument can be ap-
plied against his own sketch of the analysis of the concept ˜horse™ as well
as his analysis of ˜ought™: He does not think that zoologists are merely say-
ing that a horse is a horse and considers the claim that ˜ought™ means pro-
ductive of the best consequences to be of great philosophical importance
(as it is). So the first criticism to be made of the OQA is that, at least in
the form Moore presents it, it proves far too much, making it impossible
for analysis to be an intellectually worthwhile activity. As Moore says of
Mill™s alleged commission of the naturalistic fallacy, this problem is so ob-
vious, “it is quite wonderful how [he] failed to see it.”34
The argument poses a second problem. On a very plausible interpre-
tation of it, it just begs the question. Consider the claim “The question
“Is what we desire to desire good?” is significant.” How do we determine
that that question is significant? If we do so just by inspecting the prop-
erties denoted by the respective expressions in the question, we reverse
the epistemic order that is supposed to obtain between our determining
a question to be significant and our recognizing two properties to be dif-
ferent. Perhaps then the “argument” is really an implicit phenomeno-
logical appeal. But in that case, unless there is something that the recog-
nition of a question™s significance does to make it easier to see that the
properties are different, Moore would be advised just to have us look di-
rectly at the properties to see that they are different.
If it is possible to determine questions to be significant by some other
means than the inspection of properties, we could have a non-question-
begging argument. Interesting articles by Frank Snare and Stephen W.
Ball offer Moore a way to do just that. They rationally reconstruct his ar-
gument in a way that makes our recognition of the significance of a ques-
tion a linguistic-behavioral matter based on our ordinary understanding
of English.35 Because we are fluent speakers of English, the significance

Principia, p. 67.
Frank Snare, “The Open Question as Linguistic Test,” Ratio, Vol. XVII (1975), pp. 123“9.
Stephen W. Ball, “Reductionism in Ethics and Science: A Contemporary Look at G. E.
Moore™s Open Question Argument,” American Philosophical Quarterly, Vol. 25 ( July 1988),
pp. 197“213. Snare is more explicit than Ball that his version of the argument is not the
one Moore understood himself to be presenting.
32 g. e. moore™s ethical theory

of the relevant question is something we recognize immediately and un-
self-consciously. According to Ball, the OQA catches us “in the act, so to
speak, of actual linguistic behavior, not of meta-linguistic theorizing.”36
Snare, comparing his distinction to Ryle™s famous one between knowing-
how and knowing-that, speaks of a grammatical intuition that is “merely
a matter of having a practical knowledge . . . not a propositional knowl-
edge of the rules of grammar.”37 These distinctions are not unlike the
one Moore makes in Principia between knowledge of verbal meaning and
knowledge of analytic meaning. It is also like the distinction he makes in
“A Defence of Common Sense” and elsewhere between understanding an
expression™s meaning and knowing what an expression means, between what we
may call ordinary meaning and philosophical meaning.38 In these re-
constructions, Moore™s argument is to the best explanation of the lin-
guistic-behavioral data. It holds that our intuitive recognition of the sig-
nificance of a question is grounded in a difference in properties, which
as ordinary speakers we need not recognize in an explicit or self-
conscious manner, but which as philosophers we must explicitly come to
recognize if we are to make any philosophical progress.
Although this version of the argument does not immediately appear to
be question begging, dangers do await it. For one, there are problems
concerning the ability of ordinary speakers to recognize that a question
about a sophisticated philosophical definition is significant. What, for in-
stance, will a nonphilosopher say who is asked whether the question, “Is
that which a person qua rational agent strives for good?” is significant.
This question is difficult enough for Moore to answer. Given what he says
in the OQA, the answer has to be yes. But as we shall see, his argument
against ethical egoism, which suggests that a person who denied the iden-
tity of these notions would be guilty of a contradiction, impels him in the
other direction.39 An expression that is part of this definition does not
seem to have any ordinary meaning, leaving us with nothing to appeal to
in order to determine whether the question is significant. If this is so, the
reconstructed argument has a smaller scope than Moore™s original and
many of the most interesting definitions lie beyond its purview. If we try
to get around this fact by applying the test in these cases to philosophers™
“intuitive” understanding of terms, the obvious problem is that their un-
derstanding will differ according to what more encompassing theory they
already hold. So far then, with definitions such as this one, the recon-
struction will only have succeeded in finding a more roundabout way for
philosophers to beg the question against each other.

Ball, “A Contemporary Look,” p. 209.
Snare, “Open Question,” p. 126.
“A Defence of Common Sense,” in Philosophical Papers (London: George Allen & Unwin
Ltd., 1959), p. 37 and “The Justification of Analysis,” in Lectures on Philosophy, pp. 165“7.
See Chapters 6 and 7 for a discussion of Moore™s argument.
simplicity, indefinability, nonnaturalness 33

Assuming for the moment that ordinary speakers can always tell
whether a question is significant, we also need to wonder about the scope
of and our knowledge of the principle that is crucial to the argument. This
is the principle that a question™s significance is always due to a difference
in the properties denoted by the different expressions in the question. It
is hard to believe, as Snare recognizes, that this principle holds in all
cases.40 If it does not hold in all cases then we cannot be certain that good
is an indefinable property, since it may provide one of the exceptional
cases.41 Remember also that we are not allowed to ground our knowledge
of this principle in our ability just to inspect the respective properties de-
noted by pairs of expressions to see that they are different. Perhaps one
will try to argue that the principle can originally be based on the inspec-
tion of pairs of properties as long as none of the expressions denoting
them are “good.” Having established the principle in this way, we can then
extend it to pairs of expressions one of which is “good” without inspect-
ing the properties they denote. By doing this, we do enable the principle
to be based upon the inspection of properties. But since it is not based
on the inspection of what “good” and its purported definitions denote,
we do not beg the question. We would also need to argue that even
though we do not inspect the properties in the cases where one of the ex-
pressions is “good”, we have reason, based presumably on induction, for
believing that the principle holds in these cases as well.
But if this is the origin of the principle, one of two things would follow.
The first would be that we are able to do with “good” and the expression
purported to be its definition what we did with the expressions upon
which the principle was based: directly inspect the properties denoted by
them to see that they are different. If Moore thinks that we can do this in
this case, he would again be advised just to have us do it. The second
would be that there is some difficulty presented by the property good that
makes it not open to inspection as the properties are upon which the
principle is based. There being this difference between good and the
other properties would then make one very suspicious about the rele-
vance of the test in cases involving good. The fact that good is not open
to the same kind of inspection could plausibly be taken as evidence that
the difference between it and the other properties makes the principle
inapplicable here.
In order to avoid this general dilemma in all its guises, the linguistic
principle being appealed to must be more theoretical and methodologi-
cal in nature. The principle is not something we base upon the direct in-

“Open Question,” p. 129.
Actually, no version of the OQA can establish the indefinability of good as something cer-
tain. Moore says we must apply the test to every definition in turn. He obviously thinks
we will soon run out of plausible candidates and give up the game. Still, we can never be
certain that we have canvassed every plausible definition. See Ball, p. 209.
34 g. e. moore™s ethical theory

spection of properties or something we somehow “prove,” but something
that provides a framework for the study of linguistic behavior and the ex-
plication of the semantics of ordinary discourse. This in effect is what Ball
and Snare try to make of it. But in order to be confident about accepting
such a sophisticated principle, we need at the least to know a great deal
more about the philosophy of language and the philosophy of mind than
Moore discusses here. This supports a point we shall make a number of
times: that because of all it requires, the OQA is hardly something we have
a right to accept at the beginning of philosophical investigation. Rather
than a fully fleshed out, incontrovertible argument, Moore can only be
offering an argument sketch or hypothesis about good that the rest of the
book has the task of defending. Because he thinks of himself as offering
a more powerful argument than is really available to him, a great many
of Principia™s claims will have to be tempered “ to the benefit of his over-
all argument.
Moore fails to resolve in his own mind the tension between proving a
proposition with premises different from it and bringing forth points that
help one to have a clear perception of the proposition™s truth. There is a
passage in his discussion of the metaphysical philosophers that nicely cap-
tures this tension: “In face of this direct perception that the two questions
are distinct, no proof that they must be identical can have the slightest
value. That the proposition ˜This is good™ is thus distinct from every other
proposition was proved in Chapter 1.”42 He seems to be saying that he
proved to us that good is indefinable simply by getting us to see that it is!
The tension in Moore™s procedure lies in part in the very conservative na-
ture of his ethical project, which is much in the spirit of his later episte-
mological project of defending our commonsense knowledge of the ex-
ternal world. Argument or not, Moore is not showing us something
completely new, but something we have long known but somehow lost
sight of. Remember what he says about the connection between the nat-
uralistic fallacy and good™s nonnaturalness: We originally intuit that good
is not an existent property, but then try to make out in the roundabout
way of a definition that it is. Reflection on the significance of a question
enables us to cut through the thickets we have created to get back to what
we originally intuited and never completely forgot “ that by virtue of its
being nonnatural, good can be understood solely in its own terms.
So although Moore thinks of himself as offering an argument, what he
actually does by having us reflect on the significance of a question is shake
us out of a prejudice against accepting good™s full-fledged nonnatural-
ness. On this view, he is like the boy who tells the philosophical emperor
that he is wearing no clothes.43 He not only needs courage to make his

42 Principia, p. 126.
43 Paul Levy in G. E. Moore and the Cambridge Apostles (New York: Holt, Rhinehart and Win-
simplicity, indefinability, nonnaturalness 35

bold claim, but a genius of uncommon plainness to resist the very deep
philosophical impulse to obfuscate the simple truth of the matter. If we
read him in this way, we can capture the tension between his offering an
argument and his begging the question in the following way: There is a
sense in which he is begging the question because he is asking us finally
just to see that two properties are different. But what keeps him from
merely begging the question is the fact that he provides a new technique
to get us to recall something we used to, and at some level, still know. En-
gaging in intuitive linguistic reflection on a question™s significance takes
us out of the mode of poisonous self-consciousness that philosophy either
puts us into or results from and returns us to the purity of our original
awareness of good. Finally then, Moore™s “argument” is not an argument,
but a means for the attainment of an epiphany.
Even as they attempt to make a full-fledged argument out of Moore™s
discussion, Ball and Snare partly capture this fact in their insistence that
the reflection we perform in its name be at the level of our everyday un-
derstanding of language. Any level deeper than that and the spontaneity
is lost that is the conduit to the plain truth. The recognition of a ques-
tion™s significance sets the stage for the epiphany that finally comes with-
out the obfuscating mediation of reflection. Presumably, the crucial dif-
ference between this later awareness of good™s indefinability and the one
we previously had is that we will not let the later awareness slip away. So
Moore™s project becomes one more in a very long line of attempts to re-
turn us to, and keep us in, a state of innocence. What is striking is that in
this case, the innocence is epistemic rather than moral. (Although many
philosophers, perhaps Christian ones most prominently, will connect our
epistemic corruption to the corruption of our will.)44
Moore™s procedure here is not unlike the one he employs in “The Refu-
tation of Idealism,” published in the same year as Principia. Although
there is much sophisticated argumentation and discussion in that paper,
what is finally most striking about it is the way in which it cuts the Gor-
dian knot of a tormenting philosophical problem: “how we are to ˜get out-
side the circle of our own ideas and sensations.™” The “solution” lies in
seeing that there is no problem; just to have a sensation is to be outside
the circle.45 Moore™s arguments and discussion are aimed at getting us to
see just that. He writes, “I am suggesting that the Idealist maintains that

ston, 1979), p. 12, notes that A. J. Ayer uses this figure to describe Moore in Part of My
Life. Stuart Hampshire also uses it in “Liberator, Up to a Point,” a review of Tom Regan™s
Bloomsbury™s Prophet, New York Review of Books, March 26, 1987, p. 39.
Although we shall have a great deal to say on the subject of Moore and innocence in the
book™s last two chapters, we may note at this time that it is part of Moore™s own innocence
not to entertain deeply the possibility that our epistemic corruption is the result of some
sort of willful moral corruption.
45 “The Refutation of Idealism,” p. 27.
36 g. e. moore™s ethical theory

object and subject are necessarily connected, mainly because he fails to
see that they are distinct, that they are two at all.” Also, “My main object
has been to try to make the reader see” that consciousness and its object
are unique.46 Similarly, he “proves” to us here that good is unique by get-
ting us to see that it is.
This way of looking at the OQA fits in with some important things
Moore says about proofs in Principia. He points out that neither he nor
anyone else can prove a thing to be good or bad.47 This is not something
that should concern us, however. Our inability to prove to a madman such
obvious facts as that there is a chair rather than an elephant beside him
does not make us one whit less confident that we will not be trampled.
What bothers us about value claims is not the fact that they cannot be
proved, but the fact that there is so little agreement about them.48 But
this failure to achieve widespread agreement is due to the difficulty of the
subject, not to its having no proofs. (And on the subject of what the great-
est goods are, there is widespread agreement of a very important kind.)49
It would also seem that if we cannot prove that an object like the mad-
man™s elephant is illusory, then we also cannot prove that a property is in-
definable. What could we say to either the hallucinating or the blind that
would ensure that they see only and all of what is there? With the place
of proofs and arguments thus taken down a notch, it need not trouble
Moore that the OQA is not really an argument that proves its conclusion.
Those who are properly epistemically equipped will still receive the ben-
efit of his exercise.
But finally, we must conclude that as with all other attempts to regain
lost innocence, this one fails. If we really were to return to our original
pristine state of awareness, we could no longer have the reflective philo-
sophical understanding that is the goal of Moore™s book. In the state to
which we returned, we would take things too much for granted, would lack
the kind of wonderment that is needed to motivate a sustained philo-
sophical inquiry. And if somehow we were to become curious again, we
would again be prey to the beguilement of the intellect that made us fall
the first time. Moore has a great deal of trouble countenancing the fact
that for distinctly philosophical understanding, the danger of a special
kind of confusion must always be imminent. There are risks in self-
conscious reflection that always accompany its rewards. Given his claim
that all philosophers until Sidgwick have been under the sway of the nat-
uralistic fallacy, Moore should have realized that there is something very
deep about the kind of confusion he attempts to expose and eradicate “
it is not mere confusion to be easily overcome. There are no guarantees
that we will not fall back into it again and again.

46 47 Principia, p. 75.
Ibid., pp. 13, 45.
48 49 Ibid., pp. 188“9. See Chapters 9 and 10.
Ibid., pp. 75“6.
simplicity, indefinability, nonnaturalness 37

Moore might respond that the desirable state we achieve under his
tutelage is not quite the one we found ourselves in originally. In our new
state, we will have been chastened by the recognition that we let such an
invaluable insight slip away. Having been so chastened, we will not let it
slip away again. If we ever do find the epiphany beginning to evanesce,
we have a procedure of reflection to help us retrieve it. So we can have
the intellectual benefits of self-conscious philosophical awareness, can ac-
count for the special poignancy and depth of philosophical understand-
ing, while avoiding its pitfalls. To argue in this manner is to begin to rec-
ognize that the value of philosophical understanding is tied to the risks
of falling into a special and pernicious kind of confusion that is created
by philosophy™s refusal to take for granted what seems obvious to the
plain intellect. But if we are to treat with full seriousness this refusal to
take things for granted, we must nurture the thought that there are seri-
ous alternatives to the correct philosophical view. As Moore shows in Prin-
cipia, it is much too easy for one to be condescending and dismissive to-
ward a theory if he “knows” before studying it that it is guilty of the
naturalistic fallacy. However wrong they be even in their essentials, we
must recognize not merely the possibility, but the necessity of there being
well-conceived and compelling alternative philosophical theories. With-
out such alternatives, the refusal to take the plain awareness of value for
granted becomes inexplicable obduracy; we cannot then take account of
the achievements of a finely wrought and fully articulated philosophical
understanding. Because of the great difficulty he has in crediting such re-
flections, we shall consider in the last chapter of this book whether Moore
must finally be an opponent of philosophy, or a proponent of a recon-
figuration of it that leaves it no longer concerned with truth. These pos-
sibilities would surely be of grave concern to him, as they would seem to
leave him without a ladder to stand on.
But for now, let us bring these airy metaphilosophical musings back to
the earth of Moore™s text. In the penultimate sentence of his presenta-
tion of the OQA he writes, “Everybody is constantly aware of this notion,
although he may never become aware at all that it is different from other
notions of which he is also aware.”50 If this kind of confused awareness is
possible, then another kind of awareness, whose importance he came
later to recognize, must also be possible.51 This kind, the direct opposite
of the one he has been so concerned with, is one in which one is aware
of two properties, but mistakenly thinks of them as being only one. The
possibility of this state puts into question the assumptions of full-blown
realism, is as explanatorily powerful in the other direction, as the kind of
awareness he makes so much of. It is there to be appealed to in order to
explain why both ordinary people and philosophers think that two ques-

50 51
Principia, p. 17. Some Main Problems of Philosophy, p. 220.
38 g. e. moore™s ethical theory

tions are different when really they are the same, why they think that the
words “good” and “pleasant” denote two properties rather than one.52
The unsurprising conclusion is that there simply are no definitive
proofs for the deepest positions in ethics or any other branch of philoso-
phy. All one can do is to put forward various points for our intelligent con-
sideration, place theories in their widest possible contexts, and draw out
their deepest implications. If all this sounds vague, it cannot be helped;
the proof of the pudding is in the eating. Although one may suspect quite
early on that a philosophical theory is going deeply wrong, one must have
the patience to allow it to unfold in its entirety. Moore claims to have won
a first-round knock-out when all that philosophy allows are slow, painfully
won split decisions “ although it is true that much of the work of winning
these decisions is done by those who are brash enough to seek knock-outs.

52 Frankena, “The Naturalistic Fallacy,” pp. 112“14.
Good™s Nonnaturalness

The topic of this chapter is Moore™s claim that good is nonnatural. Our
aim is to achieve a deeper understanding of Moore™s views on the nature
of nonnaturalness and of the entire ontology within which Moore places
his theory of value, a sense of the difficulties imposed upon his concep-
tion of intrinsic value by nonnaturalness, and finally, some clues as to how
we might deal with these difficulties in a manner enriching to his theory.
To remind ourselves, according to Principia, the difference between
natural and nonnatural properties has to do with their relation to time.
Simply, natural properties exist in time while nonnatural properties do
not. Since a natural object is completely exhausted by its natural proper-
ties, no nonnatural property can be a part of a natural object. The two
kinds of property best satisfying the criteria of naturalness on this account
are: 1) the most determinate properties of colors, textures, sounds, etc.
exemplified by objects in the physical world and 2) such mental items as
feelings of pleasure and pain. When it comes to the consideration of
properties more general or abstract than these, Moore™s bare-bones
analysis of the distinction between natural and nonnatural properties
must be greatly extended and refined. Moore does a little of this work im-
plicitly in Principia and much more of it explicitly in later work, without
nearly doing all of it. The refinements he makes and the ones we make
in his name, especially as they are concerned with the nature of those very
subtle “objects,” states of affairs, show the delicacy of articulation he is
able to bring to the defense of blunt pre-philosophical conceptions, in
this case, the conception of the total objectivity and utter uniqueness of
value. Given the delicacy required by a satisfying philosophical articula-
tion of this conception, we must treat very lightly the philosophically blunt
way of stating it he chooses for his later work “ that only value properties
are nonnatural.1
In order to shed light on Moore™s views in Principia, we shall examine
two papers he wrote prior to it, “The Nature of Judgment” and “Identity,”
written in 1899 and 1901, respectively.2 The first of these papers is a wa-

Moore, “Conception,” p. 259 and “Meaning of “natural”,” pp. 591“2.
Moore, “The Nature of Judgment,” Mind n.s., 8 (April 1899), pp. 176“93. “Identity,” Pro-
ceedings of the Aristotelian Society n.s., 1 (London, 1900“1901), pp. 121“45. The essays are
40 g. e. moore™s ethical theory

tershed for him, marking his movement away from idealism to the real-
ist theories he maintained for the rest of his career. In it, he argues that
the world is composed of concepts. Concepts are possible objects of
thought and a judgment is a synthesis of concepts. Concepts are not any-
thing psychological, however, it being indifferent to their nature whether
or not anyone thinks them. In fact, our ideas are nothing different from,
but are the concepts themselves. We are thus not to understand the truth
of a judgment as being dependent on “the relations of our ideas to real-
ity.” He writes:
When, therefore, I say “This rose is red,” I am not attributing part of the content
of my idea to the rose, nor yet attributing part of the content of my ideas rose and
red together to some third subject. What I am asserting is a specific connexion of
certain concepts forming the total concept “rose” with the concepts “this” and
“now” and “red”; and the judgment is true if such a connexion is existent. Simi-
larly when I say “The chimera has three heads,” the chimera is not an idea in my
mind, nor any part of such idea. What I mean to assert is nothing about my men-
tal states, but a specific connexion of concepts. If the judgment is false, that is not
because my ideas do not correspond to reality, but because such a conjunction of
concepts is not to be found among existents.3
So before “The Refutation of Idealism,” Moore has begun his bold
move of emptying the mind of “contents” whose relation to a mind-
independent reality must always be problematic. Anticipating the later
paper, his view is that in thought the mind just grasps the mind-
independent objects themselves. The paper also makes a noteworthy dou-
ble distinction between existents and nonexistents. There is first the dis-
tinction between existent and nonexistent objects. As the nature of a con-
cept is independent of its being thought, so is it independent of there
being any existents. It would thus be wrong to think of concepts, which
are universals, as being existents or parts of existents.4 What distinguishes
the existent rose from the nonexistent chimera is the fact that the sim-
pler concepts constituting the rose stand in relation to the concepts ex-
istence and time, while those constituting the chimera do not.5 Although
he speaks of the concept “this,” which might lead one to think that his
view requires substances or bare particulars, this is not the case. Antici-
pating the views of Principia, an existent object just consists of concepts
standing in relation to other concepts, including the concept existence.
To analyze an object is to “break it down” into its constituent concepts.
Mirroring Principia™s distinction between natural and nonnatural prop-
erties, Moore also distinguishes between concepts that are capable of
standing in relation to the concept existence and ones that are not. Al-

reprinted in G. E. Moore: The Early Essays, Tom Regan, ed. (Philadelphia: Temple Univer-
sity Press, 1986), pp. 59“81, 121“47.
“The Nature of Judgment,” p. 179.
4. Ibid., p. 181. 5 Ibid., p. 179
good™s nonnaturalness 41

though he is not much concerned with nonnaturals here, he does say
some things that anticipate his Principia account of them. Since it is a con-
dition of an object™s becoming existent that it stand in relation to the con-
cept existence, existence cannot be a part of the object. Moore speaks
briefly of a sui generis relation that obtains between existents and the con-
cept existence. It is “something immediately known like red or two.”6 In
Principia, he makes room for another such sui generis relation, that which
stands between individual goods and the property good.
The second paper, “Identity,” is motivated by a change of mind on the
problem of individuation. Near the end of “The Nature of Judgment,” he
argues that objects are individuated “by the different relations in which
the common concepts stand to other concepts.”7 But in “Identity,” he ar-
gues that if the “predicates,” or properties, of different objects were iden-
tical, they could not be individuated by the relations in which they stand
to the properties of other objects. Consider two red spots, one of which
is surrounded by yellow and the other by blue. One who holds that the
red in each spot is identical will say that its two different instances are in-
dividuated by the different relations in which they stand to the yellow and
the blue. But this will not work, since on this view, “The one surrounded
by yellow is also surrounded by blue: they are not two but one, and what-
ever is true of that which is surrounded by yellow is also true of that which
is surrounded by blue.”8 One will thus be forced to say that the same thing
both is and is not surrounded by blue; he considers this a reductio of the
position. He thus concludes that the different property instances are nu-
merically different but “conceptually identical,” or exactly similar, partic-
ulars.9 This nominalistic strain is tempered, however, by the claim that
each particular stands in relation to one and only one universal, which
he likens to a Platonic form, that determines the nature of the particu-
lar.10 Two particulars, like the reds in his example, are conceptually iden-
tical if they stand in relation to the same universal. Their standing in re-
lation to the same universal is involved in the very definition of the
conceptual identity of particulars.
In this ontology, room is again made for the distinction between natu-
ral and nonnatural properties. Objects are not composed of universals
but of simple particulars, each of which stands in relation to the one and
only universal determining its nature. In fact, when we think of one thing
as having two or more properties, we are usually mistaken. An assertion
that a piece of cloth is both black and woolen, “is not to be understood
as an assertion that one individual has two predicates, but that two indi-
viduals have a certain relation.”11 Perhaps this explains why Moore can
speak in Principia of a natural property existing by itself in time. Each par-

6 7 8
Ibid., p. 181. Ibid., p. 182. “Identity,” p. 110.
9 Ibid., p. 111. 10 Ibid., p. 114. 11 Ibid., p. 127.
42 g. e. moore™s ethical theory

ticular property is logically independent of all others, even if, as does a
color, it also requires at least one other particular to be instantiated along
with it.
One of the things distinguishing particulars from universals in this on-
tology is the fact that particulars exist while universals do not. Because all
universals are nonexistent, a thing™s being a universal cannot by itself ac-
count for its being nonnatural. We need then to make a distinction be-
tween universals that determine the nature of particulars and ones that
do not. Absent such a distinction, we will not be able to explain why there
are not existent simple particular goods as there are existent simple par-
ticular reds and sweets. In the paper, for a different reason, Moore makes
a distinction between kinds of universals that can also be applied to this
one. It is obvious that a complex has properties that none of the con-
stituents of the complex have.12 When we ascribe one of these properties
to the entire complex, we are making a different assertion than that a re-
lation or set of relations obtains among the different particulars com-
prising the complex. Therefore, these properties are not parts of the ob-
ject in the way that simple particular properties are. They thus satisfy one
of Principia™s criteria of nonnaturalness. We can also see that simple par-
ticulars have properties that do not determine their nature. If, to use one
of Moore™s own examples, a complex has the property of being so many,
then each particular has the property of being one.13 This property plays
little or no role in the determination of an object™s nature. While it is true
of any particular that it would not be the thing it is were it not one, its be-
ing one does not determine which one it is. Moore calls such nondeter-
mining properties class-concepts.14
If we apply this distinction to good, we get the following account: From
the fact that good is not a property that determines the nature of any-
thing, it follows that no good thing is conceptually identical with the
property good. This points to a way in which one can commit the natu-
ralistic fallacy within this ontology. (The fallacy™s great versatility enables
it to be committed within every ontology!) To commit the fallacy is to
place good in the wrong category of universal, to think of it as one of the
universals that does determine the nature of the simple particular things
exemplifying it. Once it is placed within the wrong category, the tendency
toward moral monism, the view that only one kind of thing is good, be-
comes irresistible. In this ontology, the simplicity of pleasure makes he-
donism the version of monism most likely to be irresistible.
Another class-concept Moore mentions is the determinable red.15 We
might be tempted to think that there is some relation like conceptual
identity holding between different determinate universal reds that is to be
analyzed in terms of the relation in which each of them stands to the

12 13 14 15
Ibid., p. 126. Ibid., p. 117. Ibid., p. 124.
good™s nonnaturalness 43

higher-order determinable red. But Moore has two reasons, which we
have already encountered in Principia, for denying this. (Although the
implicit appeal to a second kind of definition applies a chastening ele-
ment to the first of those reasons in Principia.) The first is that he con-
siders universals to be “mere points of difference.”16 The second is his
fear that such a view leads to extravagant idealist commitments to “con-
crete” or “self-differentiating” universals and to things having “identity in
difference.”17 To anticipate a concern on the horizon, on this model, all
determinables are nonnatural. This makes the category of class-concepts
less than a perfect home for good. He tells us later that he never thought
of good as a determinable and places determinables explicitly into the class
of natural properties.18 But at this time, he has not yet turned his thoughts
explicitly toward the ontology of value and thus has not yet thought about
how to distinguish good from natural determinable properties.
It is probably in “Identity” that Moore™s inclination to see analysis as be-
ing required always to go downward toward the most determinate natu-
ral properties of a thing is at its strongest. The unchastened view that uni-
versals are mere points of difference seems to prevent them from having
any features open to analysis in terms of more general properties. Also
pointing analysis down is the claim that there are not single objects hav-
ing two different properties such as being black and being woolen, but
rather two different objects standing in relation to each other. Both of
these views hint that it is an arbitrary matter what unities the mind finds
among, or better, imposes on, the world™s simple objects. The first of
these views does not seem to allow that there is anything more in com-
mon between an instance of red and an instance of green than there is
between an instance of red and an instance of sweet. The view that what
we ordinarily consider to be single objects really are not seems to prevent
the object consisting of the particular blackness and particular wooliness
of Moore™s example from being any more of a unity than the “object” con-
sisting of that same blackness and a particular cottony quality in a pink
Fortunately, in his work after “Identity,” Moore never allows the down-
ward conception of analysis to have complete ascendancy over the up-
ward. In Principia, remember, the second sense of definition he appeals
to is one that goes up to the more general. He also recognizes an obliga-
tion to make sense of the fact that it is not an arbitrary matter which prop-
erties are seen by the mind as forming unified wholes. His view suggests
that there are natural complexes, complexes that really do have some kind
of unity within their complexity. Such complexes as animality are, at the
requisite levels of abstraction, the objects of the various sciences. To al-

16 17 Ibid., p. 125.
Ibid., p. 120.
Moore, “Meaning of “natural”,” p. 583.
44 g. e. moore™s ethical theory

low that it is not arbitrary which objects are seen as forming unified
wholes, it is important that we let the objects at different levels of analy-
sis stand on their own feet, that we not seek fully to reduce one level of ob-
ject to another.19 This is to suggest that objects instantiate more or less
fully a number of different natures, that there are various degrees of
looseness without arbitrariness in the identity of objects at different lev-
els of analysis, and also in the relations obtaining among the particular
qualities that combine with each other to form objects.
Not the least important consequence of allowing a spirit of ontological
expansiveness to accompany the spirit of analytical rigor is the way in
which it can be used to enrich Moore™s moral epistemology. His overall
theory is at its best when it allows the world to contain many more dif-
ferent kinds of natural goods, of which we start to become aware very early
in life, than he officially recognizes. Perhaps, for instance, we grasp the
complex object health and its goodness, while being quite ignorant of the
exact properties constituting health. Such a view makes the rough and
ready, but nevertheless very deep, agreements in the opinions of human
beings about what things are good much more a matter of insight than
of indoctrination.

Later Refinements and Problems
“The Conception of Intrinsic Value” was written in the teens and first pub-
lished in 1922.20 In this great paper, Moore extends and refines his Prin-
cipia views on the distinction between natural and nonnatural properties.
He then refines those views in response to criticism by C. D. Broad;
Broad™s criticism and Moore™s response are both found in The Philosophy
of G. E. Moore, published in 1942. The greater subtlety of Moore™s later
views does not prevent them from continuing in the vein of Principia. In
fact, he uses their greater subtlety to sharpen the distinction between nat-
ural and nonnatural properties and asks the distinction to carry even
more weight than it does in Principia. He winds up holding that only value
properties are nonnatural and makes nonnaturalness solely responsible
for value™s being an intrinsic feature of the world.21
Before we examine the details of “Conception,” let us clear up a ter-
minological difficulty that Moore points to in his reply to Broad.22 Moore
claims that the value of a thing is a property whose instantiation depends
solely on its intrinsic nature. Thus it is a property that is intrinsic to the
thing. But because it differs from those properties constituting the thing™s

Butchvarov, Skepticism in Ethics, p. 60.
Moore, Preface to Philosophical Studies, p. vii.
Moore, “Conception,” p. 259. Although Moore uses beauty as his example of an intrin-
sic value property in this paper, we shall speak of the property good.
22 Moore, “Meaning of “natural”,” pp. 583“5.
good™s nonnaturalness 45

nature, it is not an intrinsic property. He is thus forced to say that although
value is a property and is intrinsic, it is not an intrinsic property! In his
reply to Broad, in order to avoid confusion, he distinguishes between nat-
ural intrinsic properties and nonnatural intrinsic value properties. Al-
though we shall follow him in this change as much as possible, we shall
occasionally have to rely on his original, “awkward” terminology.
We begin with Moore™s claim that value depends solely on the natural
intrinsic properties of a thing. For this to be the case, two theses must be
(1) that it is impossible for what is strictly one and the same thing to possess that kind
of value at one time, or in one set of circumstances, and not to possess it at an-
other; and equally impossible for it to possess it in one degree at one time, or in
one set of circumstances, and to possess it in a different degree at another, or in
a different set. . . . (2) . . . if a given thing possesses any kind of intrinsic value in
a certain degree, then . . . anything exactly like it, must, under all circumstances,
possess it in exactly the same degree. Or to put it in the corresponding negative
form: It is impossible that of two exactly similar things one should possess it and
the other not, or that one should possess it in one degree, and the other in a dif-
ferent one.23
To understand these theses correctly, we must have a sure grasp of their
modal terms. Moore insists that the impossibility being invoked is not
merely causal.24 If it were, the theses would fail to preclude certain kinds
of subjectivist analyses. It is causally possible, for instance, for two exactly
similar things to elicit exactly similar responses in anyone who perceives
them in exactly similar ways. This would allow for an analysis of value in
terms of perceivers™ attitudes. The requisite impossibility must be such as
to guarantee that things having intrinsic value have the same amount of
it no matter what any perceiver™s attitude toward them happens to be.
But the proper rendering of the modality of these conditions alone
does not insure the nonnaturalness of value. These conditions could ob-
tain even if good were an intrinsic natural property, for example, the
property of being a state of pleasure. In order to preclude naturalistic
analyses in terms of pleasure or any other natural property, another the-
sis must be maintained: Although it is solely dependent on the intrinsic
natural properties of an object for its instantiation, goodness is not one
of those properties. Rather, a special relation of dependence obtains be-
tween the natural properties of an object and its goodness. According to
Moore, this relation serves to make value properties utterly unique; he
“cannot think of any other predicate which resembles them in respect of
the fact, that although not itself intrinsic, it shares with intrinsic proper-
ties the characteristic of depending solely on the intrinsic nature of what
possesses it.”25

23 24 25
Moore, “Conception,” pp. 260“1. Ibid., pp. 267“8. Ibid., p. 273.
46 g. e. moore™s ethical theory

In order to see how the fact of intrinsic value™s being different from but
completely dependent on natural intrinsic properties precludes different
sorts of naturalistic theories, consider two different examples of such the-
ories: 1) “The assertion “A is good” means “A is pleasant” and 2) “The as-
sertion “A is good” means A is a state of pleasure.”26 The first type of the-
ory is precluded by the fact that while a thing™s intrinsic value is solely
dependent on its intrinsic natural properties, its pleasantness is not. The
pleasantness of a thing is at least partly dependent on the attitudes of the
people experiencing it. Some people find a thing pleasant “ like it “ and
others do not. On the second type of theory, a state of pleasure is an in-
trinsic natural property of a different, more inclusive object than the ob-
ject of which the pleasantness of 1) is an extrinsic property. To combat this
type of theory, it must be maintained that its goodness is not an intrinsic
natural property of that object. But since its being a state of pleasure is an
intrinsic natural property of the more inclusive object, it is allowed that the
goodness of that object is dependent on the state™s being one of pleasure.
We can flesh out our understanding of the way in which the two cru-
cial features of nonnaturalness keep one from conflating being good with
being pleasant or being a state of pleasure by examining an argument of
Broad™s and Moore™s response to it. Broad claims that the pleasantness of
a taste is always “derivative” upon certain of its “non-hedonic” properties,
such as its sweetness. As he puts it, there must always be an answer in
pleasure-neutral terms to the question, “What makes the thing pleasant?”
In like manner, he finds goodness to be derivative. The goodness of a cer-
tain experience, for instance, is dependent upon its being a sorrowfully
tinged awareness of another person™s distress. But the pleasantness is
supposed to be natural and the goodness not. He writes:
Now Moore counts pleasantness as a natural characteristic. If he is right in doing
so, it is impossible to identify the nonnatural characteristics of a thing with the
derivative sub-class of those of its characteristics which depend solely on its in-
trinsic nature. For by that criterion pleasantness would be a nonnatural charac-
teristic just as much as goodness.27
Moore responds that Broad™s criticism rests on a failure to distinguish
between two senses of the word “pleasant.” Each of these senses applies
to a different kind of thing. When we become clear about what kind of
thing each applies to, we see that in both its senses, pleasantness is quite
a different kind of property than goodness. According to the first sense,
pleasantness is an extrinsic property. Therefore, “the fact that it is “natu-
ral” can have no tendency to show that there are any natural derivative
intrinsic properties.”28 As an example, Moore considers numerically dif-

C. D. Broad, “Moore™s Ethical Doctrines,” in The Philosophy of G. E. Moore, p. 61.
28 “Meaning of “natural”,” p. 588.
good™s nonnaturalness 47

ferent but qualitatively identical tastings of caviar. To some people such
tastings are pleasant, while to others they are not. Since, pace Berkeley,
the bare taste is the same whether one finds it to be pleasant or unpleas-
ant, neither its pleasantness nor its unpleasantness is an intrinsic prop-
erty of the tasting of caviar. Thus as Moore says, the pleasantness is de-
rivative and extrinsic and on a different plane than good, which is
derivative and intrinsic.
When we wish to account for the contribution to goodness that the
pleasantness of things makes, we must turn to the larger sorts of wholes
to which the second sense of “pleasant” applies. Continuing with Moore™s
example, “pleasant” applies in this instance “not only [to] an experience
of tasting caviare but also an experience of feeling pleased with the taste.”
Of this larger object, the property of being pleasant is not derivative. As
Moore points out:
A state of things which can be properly described by stating that it is a state of
things in which some person is both tasting the taste of caviare and being pleased
with the taste, cannot be exactly like a state of things . . . in which some person is
tasting the taste of caviare but not feeling pleased with the taste.29
Since the second sense of “pleasant” names a natural, nonderivative
property of a more inclusive whole, its pleasantness is one of its good-
making properties. (To remind the reader again, we shall say more later
about such larger wholes, “states of things” in Moore™s terminology,
“states of affairs” in others™.)
Moore admits in “Conception” to being unable to say what exactly it is
that distinguishes natural from nonnatural properties. All he can do is
give a “vague expression” of the difference. He says that natural proper-
ties “describe” an object while nonnatural properties do not:
If you could enumerate all the [natural] intrinsic properties a given thing pos-
sessed, you would have given a complete description of it, and would not need to
mention any [nonnatural] predicates of value it possessed; whereas no descrip-
tion of a given thing could be complete which omitted any [natural] intrinsic prop-
In his reply to Broad, Moore notes that this account is not quite accu-
rate as it stands.31 Although they are natural intrinsic properties, an
object™s determinable properties need not be mentioned in a complete
description of it. Since these properties follow from the object™s deter-
minate properties, mentioning them in addition to the determinate
properties describes the object no further. So the natural intrinsic prop-
erties consist of the most determinate properties that describe the object
and the determinable properties entailed by them. This characterization
makes it impossible for good to be a determinable and in this discussion

29 30 “Conception,” p. 274.
Ibid., p. 589.
31 “Meaning of “natural”,” pp. 586“7.
48 g. e. moore™s ethical theory

Moore claims never to have thought of good as a determinable. By ex-
plicitly disallowing this possibility, he makes it obvious that he requires a
special and unique relation of necessity to obtain between a thing™s nat-
ural properties and its goodness.
Clearly, Moore has brought great concentration, care, and subtlety to
the subject of nonnaturalness in both “Conception” and his reply to
Broad. As his discussion of the example of the enjoyment of caviar shows,
he is considering “parts” and “objects” whose analyses must be quite a bit
more nuanced than those that apply to brute physical objects. One might
even think that we should completely abandon talk of parts and objects
in coming to grips with his views. In support of this recommendation is
the fact that Moore says in his reply to Broad that he no longer thinks of
its brownness and roundness as parts of a penny.32 So if we are to con-
tinue to make use of the notion of parts, we must do so with great care.
Despite the greater subtlety of his later account and his unhappiness
with his earlier way of making the distinction, Moore still deeply retains
the spirit of Principia. The distinction between the tasting of caviar and
the enjoyment of the tasting of it is similar to the distinction he makes in
Principia, as part of his argument against the view that pleasure is the sole
good, between pleasure and the consciousness of pleasure.33 The greater
flexibility of his later work lets us give a more exact characterization of
the distinction than could be given in Principia. Enjoyment is a distinct
element of an experience, being something that can be identified by itself,
even though, because it must always be the enjoyment of something, it
cannot exist by itself. It counts as natural because it has duration and be-
longs to wholes that can exist by themselves.
Even with the new characterization, the distinction between natural
and nonnatural properties is at its starkest and easiest to grasp at the level
of the most determinate properties. But once we start to consider subtler
sorts of properties, it is not so obvious that there is as sharp a distinction
between natural and nonnatural properties as Moore thinks. To start, is
good really so unlike determinable properties that it must be placed in a
category utterly separate from them?34 Questions posed in terms of de-
scriptive and nondescriptive predicates seem no less trenchant. To cite
something he realizes but does not make enough of, there are many dif-
ferent ways of describing.35 Are there not then degrees of descriptive-
ness? If the most important and interesting task of philosophy is “To give
a general description of the whole of the Universe,”36 is there not a sense

32 33 Principia, pp. 87“91.
Ibid., p. 582.
Butchvarov, “That Simple, Indefinable, Non-natural Property Good,” pp. 58“62, Skepti-
cism in Ethics, pp. 62“3.
“Meaning of “natural”,” pp. 591“2.
36 Moore, Some Main Problems of Philosophy, p. 1.
good™s nonnaturalness 49

in which we are describing the universe when we say that there is intrinsic
value in it?37 Such a sense would provide Moore with the clearest way of
distinguishing between his view and the views of such opponents as emo-
tivists and prescriptivists, who also maintain that to say of something that
it is good is not to describe it. These and other philosophers like them
might feel less hostility toward allowing good into their ontology if they
see that a richer typology of properties is available once they take their
eyes off the paradigm natural properties. Even if we continue with Moore
to consider value to be utterly unique, there is great heuristic value in re-
membering his Principia distinction according to which many properties
are nonnatural. Showing good™s likenesses to some of these other prop-
erties helps make a belief in it seem less “superstitious.”38
It is not just from the global perspective that the ascription of goodness
can be seen to be descriptive. Consider an example from art. Suppose
that one person A sees that a piece of sculpture is beautiful and therefore
good while another person B does not. Since the two can agree on the
most cold-blooded inventory of the thing™s parts and arrangements, there
is the Moorean sense of description according to which the dispute about
its goodness is not descriptive. But there are other ways of describing and
seeing, ways familiar to gestalt psychologists, for instance, upon which they
do not agree. This kind of seeing is not a matter of espying different parts,
but a matter of seeing the same parts in a different way (seeing that the
duck is also a rabbit). When A tries to get B to see that the sculpture is a
good piece of work, she attempts to get her to “see the thing in a new
light” and not merely to get her to “change her attitude” toward it. If A
succeeds, B sees the grace or the power of the piece she had not seen be-
fore. The sculpture makes sense or raises questions for B when before it
had not. In learning to see the piece in this way, B is learning how it
achieves (its) good.
In most cases, the perception of goodness is dialectical and expansive.
We cannot get an exhaustive understanding of a thing™s goodness “ nor
even of the thing “ by focusing on it alone, but only by comparing it to
other things and placing it within the classes of things into which it fits
most comfortably.39 This makes the investigation and description of
things from which their goodness is seen to follow wide-ranging and the-
oretical. We do not realize just how wide-ranging and deeply enmeshed
within theory these things are because we begin to learn the theories so
early in our lives and so deeply that we take them for granted. (Our train-
ing in aesthetic theory begins even before crayons and bedtime stories.)

37 38 “Conception,” pp. 258“9.
Ibid., pp. 26“7.
Butchvarov, Skepticism in Ethics, pp. 69“70. In “The Limits of Ontological Analysis,” in The
Ontological Turn, E. D. Klemke and Moltke Gram, eds. (Iowa City: University of Iowa Press,
1974), pp. 3“37, Butchvarov claims that all understanding is comparative, going beyond
the bare description of things and requiring classification.
50 g. e. moore™s ethical theory

It only seems to us now, in some cases, that we just “see” the goodness of
things as we see the yellowness of them. Moore™s own account in Principia
of the ascription of beauty to a thing is at a very great remove from the
widely accepted caricature of his view on the perception of value accord-
ing to which we need only to perform the quickest inventory of a thing™s
most determinate natural properties before rendering a judgment on
whether goodness emanates from it.40 But perhaps because he never again
engages in the same detail of aesthetic discussion, he does not consider in
his later career how far his views on the ascription of goodness or beauty
can get from the starkly drawn distinction between descriptive and non-
descriptive predicates that his official conception of analysis appeals to.
In Principia remember, Moore defines the beautiful as “that of which
the admiring contemplation is good in itself.” He then elaborates on this
To assert that a thing is beautiful is to assert that the cognition of it is an essential
element in one of the intrinsically valuable wholes we have been discussing; so
that the question, whether it is truly beautiful or not, depends upon the objective
question whether the whole in question is or is not truly good, and does not de-
pend upon the question whether it would or would not excite particular feelings
in particular persons.41
By the lights of this definition, when we ascribe beauty to a thing, we do
more than say that it has a certain property that follows from the prop-
erties mentioned in the bare description of it. The ascription of beauty
ranges beyond the individual thing to which we are ascribing the beauty
and places it explicitly into one or more classes of thing. It is thus intrin-
sically classificatory and comparative. Even though we need not think ex-
plicitly of these classes of objects when we ascribe beauty to something,
we have some vague sense of them in mind; the membership of the thing
in such classes is part of what we mean. Although it may be in spite of ini-
tial appearances, we do have part of our eye on those classes and not just
on the thing to which we are ascribing beauty.
Notice also how subtle and complex Moore™s view on the ascription of
beauty in such cases is. He claims that “to say that a thing is beautiful is
to say, not indeed that it is itself good, but that it is a necessary element in
something which is: to prove that a thing is truly beautiful is to prove that
a whole, to which it bears a particular relation as a part, is truly good.”42
But in order to decide whether the larger whole consisting of the cogni-
tion of the object is good, we do not directly examine it, but instead the
aesthetic object that is a part of it and other specific and general aesthetic
objects to which the aesthetic object is related. In fact, we can only create
the intrinsically valuable whole, the appreciation of the aesthetic object,

We shall take issue with this caricature again in Chapter 4.
41 42
Principia, p. 201. Ibid., p. 202.
good™s nonnaturalness 51

by studying one of its parts, the aesthetic object. If we are still studying
the part, the larger whole is obviously not yet completed for examination.
There appears to be a most interesting similarity between this procedure
of creative discovery, wherein we create an intrinsically valuable larger ob-
ject by looking for value in a smaller object, and the procedure Socrates
recommends for living the best kind of life.
Moore also speaks of “proving” that a whole to which the part bears a
particular relation is good. But since he has already told us that we can-
not really prove that a thing is good, he must be considering a proof in
some weaker sense of the term. To prove that a thing is good in this sense
is to show that it is a member of the class of acts of aesthetic appreciation,
whose goodness has been established in the preceding pages of Principia.
Showing something controversial to be an act of aesthetic appreciation
can get quite complicated. The long and the short of it is that we show
that a controversial object is sufficiently like things uncontroversially be-
longing to that class as also to warrant inclusion in it. This introduces a
conservative element to the methodology of such proofs. But that con-
servative element is offset by the fact that a successful proof expands the
class of objects we recognize as being worthy of appreciation. Proving
something to be aesthetically valuable will in certain cases be long and re-
visionary rather than short and conservative.
For instance, to rehearse what has been going on in American house-
holds and in popular and learned journals for over two decades, one
shows to noninitiates that listening to rap is sufficiently similar to other
kinds of aesthetic experience as also to warrant classification as an aes-
thetic experience. One way to do that is to show that rap is sufficiently
similar to one or more of the members of the class of African-American
musical objects, rhythm-and-blues, or reggae, say, as also to be included
in it. One may also compare rap to other forms of music and other
declamatory and poetic forms. But keeping our eye just on the defense
of rap as being similar to other forms of African-American music, one
who is challenged about them can defend their merits by pointing out
their likenesses to other kinds of music. (One may also feel that their mu-
sicality is so obvious as to need no argument.) One can then defend the
forms that provided the second comparison class by comparing them to
others. This process can go on for quite some time before something is
simply accepted as musical.
While the defense has so far been conservative, there is also a revi-
sionary element in the proof of the aesthetic worthiness of rap. In order
to show that rap is a distinct art form, that it has its own kind of beauty,43 at
some point one must highlight the features that make it unique. Because
rap does have its own unique beauty, before we had ever listened to it, we

52 g. e. moore™s ethical theory

had not known that something could be beautiful in that way. If we have
been successfully prepared to hear rap™s beauty, our study of the more
specific will have enriched our understanding of the more general. The
inclusion as music of something so close to the spoken word and so heav-
ily rhythmic deepens our appreciation of the source of music in the con-
versation and movement of everyday life. Our receptivity to beauty grows
along with our admiration for human inventiveness as we are exposed to
this new kind of beauty.
To repeat, when we appreciate something aesthetically, we do not just
study and learn about it. We look at its relations to many other things and
their relations to many other things and learn about them all. By following
Moore himself, we have strayed very far from the paradigm he is supposed
to blindly accept according to which we need only note the most determi-
nate properties of a thing before deciding whether the nonnatural prop-
erty of beauty or goodness follows from it. There is no reason to think that
there is not a similar complex dynamic involved in the study of Moore™s
other great good, friendship, or any other goods we may happen to find.

Final Refinements
Our final discussion concerns the possibility of maintaining of certain
other kinds of “objects,” most importantly of those known as states of af-
fairs, that their intrinsic value depends solely on their intrinsic nature.
Maintaining this view will require us to stretch even wider the conception
of objecthood that Moore propounds in “The Conception of Intrinsic
Value” and his reply to Broad. One of the most interesting features of
Moore™s work in this area (as in others) is its ability to join a spirit of ex-
pansiveness “ more probably than he recognizes “ to a spirit of analytical
austerity. Although we must remain respectful of certain of its limitations,
which at the levels of abstraction being considered are not exactly weak-
nesses, it does appear that a reasonable Moorean explication can be
given, for any object that can plausibly be considered to have intrinsic
value, of the distinction between the property it exemplifies of being in-
trinsically valuable and the properties it has as a part of its intrinsic na-
ture that make it intrinsically valuable.
Let us begin by considering the nature of things much like art objects:
tools and artifacts “ knives, can openers, cars, and so on. Borrowing an
example from Philippa Foot, we can envision that a thing found in an-
other culture, with spatio-temporal properties identical to the properties
had by a thing our culture uses for cutting, is there a decorative object,
not a knife.44 If this is so, then any intrinsic value that such things as

Philippa Foot, “Goodness and Choice,” Virtues and Vices (Berkeley: University of Califor-
nia Press, 1978), p. 134.
good™s nonnaturalness 53

knives have qua tool, or any intrinsic value they contribute to as parts of
larger wholes, cannot be due solely to their simple determinate spatio-
temporal properties and the determinable properties entailed by them.
Any such value is heavily, perhaps solely, dependent upon their having a
function and their performing it well. There is nothing whose function
is determined by its spatio-temporal characteristics alone.
One response to this problem is that these things have only instrumen-
tal value. Since the instrumental value of a thing may vary depending on
whether it is used for cutting or for decorating, its intrinsic natural prop-
erties need not go beyond the spatio-temporal ones. It will be difficult for
Moore to accept this conclusion, however. Such objects are beautiful ac-
cording to the Principia definition that what is beautiful is that of which
the admiring contemplation is intrinsically good. In most instances, these
things can only be rightly admired by one who knows what they are for.
Even if one rejects Moore™s argument that the mere existence of beauti-
ful objects has intrinsic value, it remains that these objects, being worthy
of admiration, are capable of being intrinsic parts of larger wholes having
intrinsic value.45 We must then have a more nuanced account of their na-
tures, an account that makes their use a “part” of them. To help develop
this conception, we again remind ourselves to be respectful of the fact that
the different higher-level natures a thing instantiates are not completely
reducible to its lower ones. Although it is true that at one level of analysis
a thing just is its brute spatio-temporal properties, qua artifact, it is more
than that. In the full description of a thing as an artifact, one must appeal
to aforementioned facts about the culture in which it was made and the
relevant intentions of its maker(s). Once having brought these features
of a thing into play, we can say of our knife and their decorative object
that even though their determinate physical spatio-temporal are identi-
cal, they do not belong to the same kind and hence are not identical.
When put so strongly, this reply sounds odd. If two things are com-
posed of the very same determinate properties, the same stuff, how can
they not be the same? One replies by repeating that there is more to be-
longing to a certain kind than having the right stuff. The objector is for-
getting that analysis must proceed at different levels, must assume dif-
ferent perspectives. From the right perspective, it is no longer odd to say
that for a thing to be a certain kind of artifact, certain social facts must
obtain. When we say that these properties are a part of what it is to be a
knife, we must remember that we are now using a very abstract and at-
tenuated sense of the term “part.”46 To make this thought more plausi-

For Moore™s “Beautiful World Argument,” see Principia, pp. 83“5.
A remark of Butchvarov™s in “The Limits of Ontological Analysis,” pp. 26“7, is apposite:
“The analytical understanding regards the features and circumstances in terms of which
it seeks to understand the nature of the object as possibly only analogous to parts of the
object, not necessarily as literally parts of the objects.”
54 g. e. moore™s ethical theory

ble, consider the reaction of archaeologists to a cultural object they dis-
cover of whose function they are ignorant. They cannot classify the ob-
ject, say what it is, until they know its function. Does its function not then
seem to be a “part” of its nature?
To bring facts about functions into the analysis of these objects, it will
be helpful to make a distinction similar to one we make with acts of en-
joyment and with beliefs. Even though we cannot describe the nature of
a particular act of believing without saying what it is about, we recognize
that what it is about does not affect its nature qua belief. In some manner
such as this does a thing™s being a knife depend upon social facts without
those facts literally being part of the physical thing that is the knife. It is
conceded that Moore might not have been happy to see his account go
so far in this direction. Remember the contempt he expresses for the
claim that its being a part of the body is part of the analysis of being an
arm.47 Still, his view seems not only to allow, but to require this kind of
expansion if we are to continue to maintain that the intrinsic value of a
thing is solely dependent on its intrinsic natural properties.
Perhaps it is more obvious with functional objects than with others that
we will have to move some distance away from one of Moore™s guiding
conceptions on the nature of good™s dependency on the natural proper-
ties of objects. For something to be a functional object, it is enough that
its makers have had some very general intentions about what it was sup-
posed to do. Any more precise intentions concerning the object are thus
irrelevant to its nature qua functional object. Insofar as an object™s value
is dependent on how well it performs that general function, it is possible
for its value not to depend on its most specific spatio-temporal proper-
ties. Mark McGwire™s baseball bat may be equally lethal whether it is col-
ored black or white, or even whether it weighs thirty-two or thirty-three
ounces. Unsurprisingly, given the similarity of functional objects to art
objects, this point also seems to hold for art objects. Part of the beauty of
a painting, for instance, can lie in a contrast that could have been brought
about by any number of different specific colors. So despite what Moore
says in Principia, the most specific properties of an art object need not be
“essential to its beauty.”48 A little more thought suggests this point to hold
across the board for value. The goodness of a certain act of friendship de-
pends on its being an act of encouragenent, not on its being a pat on the
back or the saying of a particular word. Anyone who worries that value
does not then touch all the way down to the bottom of things is reminded
here that the relation between lower and higher levels of objects is not
fully reductive and also of Moore™s unofficial conception of definition
that ties ultimacy in natural properties to general properties rather than

47 48
Principia, pp. 31“3. Ibid., p. 202.
good™s nonnaturalness 55

A consideration of artifacts brings us now to those most important and
abstract kinds of objects, states of affairs.49 Immediately upon moving to
states of affairs, we can say that the state of affairs that consists of a thing™s
existing while having been made in a certain culture by people with cer-
tain intentions is different from the state of affairs that consists of a phys-
ically identical thing™s existing while having been made in a different cul-
ture by people with different intentions. This suggests a principle that
only seems paradoxical at first glance: States of affairs containing differ-
ent physically identical objects may have different natures and hence dif-
ferent amounts of intrinsic value.
Perhaps the thought that all value attaches to states of affairs is found
in Principia and other of Moore™s works without much being made of it.
At one point in Principia, Moore says, against the metaphysical philoso-
phers who find value only in what is eternal, “But when we assert that a
thing is good, what we mean is that its existence or reality is good.”50 The
method of isolation for determining the amount of an object™s intrinsic
value also suggests that what it is for an object to have intrinsic value is
that the state of affairs consisting of the object™s existing all by itself has
intrinsic value. Later, in Ethics, Moore offers as the analysis of the claim
that A is intrinsically better than B, “˜it would be better that A exist quite
alone than that B exist quite alone.™”51
The claim that all value resides in states of affairs might leave one un-
comfortable for a number of reasons. One might object on phenome-
nological grounds that we find value in the thing itself and not just in its
existence. When we wonder whether a thing is intrinsically good, we con-
sider it and not its existence. Recalling for a moment the ontology of “The
Nature of Judgment,” the nature of a concept is prior to and indepen-
dent of the concept existence “ the chimera is the same thing whether it
exists or not. “The Refutation of Idealism” also seems to require that the
mind engage directly with nonexistent things. One well-known response
to views such as these is that to think that there is a significant difference
between thinking of an object and thinking of an object™s existence is to
fall prey to an act of philosophical sleight of hand. Kant™s famous dictum
that existence is not a predicate (that it does not describe) is an attempt to
head off the dangers that come from indulging in such lines of thought.
Still, insofar as we are concerned specifically with Moore, we are likely to
have to allow a great deal of room for Meinongian views, as he holds that
it is in the nature of an art object to be unreal.52
It should be mentioned here that it might be the case that certain kinds
of facts are relevant to the determination of the value of some states of
affairs, but irrelevant to the determination of the value of others seem-

49 50
See Baldwin, G. E. Moore, p. 73. Principia, p. 120.
51 Ethics, p. 39. 52 See Chapter 10.
56 g. e. moore™s ethical theory

ingly rather like them. Consider, for instance, how a knowledge of the
state of an artist during the time he created a work might sometimes be
relevant to the value of our appreciation of his work. The fact that the
artist composed a great symphony despite suffering a loss of hearing
might make his achievement worthier of appreciation than it would oth-
erwise be. The state of affairs consisting of our appreciation of the work
while knowing of the creator™s struggles with his disability might then
have greater value than the state consisting of our appreciation of the
same work while knowing that it had not been created by someone so af-
flicted. Conflicting with this thought is our inclination to say that just so
long as there are the relevant general intentions on the part of the com-
poser, the value of the work and its existence and our appreciation of it
are independent of other facts about him. Grown-ups must accept that
trying hard does not guarantee a thing™s goodness (although some value
might, pace Moore, attach to the trying).53 One might also object that the
view puts us on a dangerous slippery slope. Is a limitation of talent or even
laziness and other kinds of moral or temperamental deficiencies an af-
fliction the overcoming of which gives greater value to certain works or
actions and our appreciation of them?
Before responding that the answer to this question is obviously no, con-
sider the grandeur that Malcolm Lowry™s struggles with alcoholism give
to Under the Volcano or the depth that is added to Miles Davis™s trumpet
playing by his having to come to terms with a less than breathtaking tech-
nique. Similar questions pose well-known theoretical and practical prob-
lems in the more distinctly moral sphere. Remember Kant™s worries about
the moral worth of actions done by people who are temperamentally in-
clined to do the right thing. There is certainly something moving about
people struggling with and rising above their limitations. Still, it is easy to
point to other cases where we are inclined to say just the opposite of what
we are inclined to say in these cases. In art, we find something thrilling
about work, like that of the young Elvis Presley, that is the expression of
easy, fecund genius. Morally, we consider a naturally good nature to be
very good “ and again, pace Moore, not just as a means. Likely, the sug-
gestion we make in the book™s last few pages that there are not just pure
art objects that we appreciate but continua of art-historical objects, would,
along with a similar point about objects of more distinctly moral evalua-
tion, be quite useful in an extended discussion of these matters. Some of
the different claims that seem true but inconsistent if about the same
things, are actually about different things. This, of course, still leaves
open the question how much each of these different kinds of things con-
tributes to the value instantiated by the world taken as a whole.
Something we need to consider in dealing with these and related issues

53 Ibid., p. 176.
good™s nonnaturalness 57


. 2
( 8)