Moore
Moore
Moore
Graham Priest
March 8, 2017
1
an enticing and prevalent misconception about the history of
metaphysics: that its principle value lies in its indicating voices
of yore that can be heard as participating in contemporary meta-
physical discussions. On the contrary, its principle value lies
in its indicating voices of yore that cannot be heard as partic-
ipating in contemporary metaphysical discussion. It indicates
voices that challenge whatever presuppositions make contempo-
rary metaphysical discussions possible.
In the same way, the best way to understand one’s own culture is to under-
stand others.
The central concern of Moore’s book is, as the title says, metaphysics,
though Moore’s definition of ‘metaphysics’ is a somewhat unusual one (EMM,
p. 1):
Metaphysics is the most general attempt to make sense of things.
This sounds to me more like philosophy itself than metaphysics; but I have no
desire to quibble about words. The topic is clearly a profound and important
one; and whatever one calls it, making sense is the lens through which the
discussion in the book is focussed.
If one is of a certain post-modern persuasion, one might think that the
enterprise, so defined, is a forlorn one. There is no “master-narrative” which
can satisfy such a global attempt. Moore is sensitive to the fact that ways
of making sense are a plurality. They form a united enterprise for all that.
Indeed, to suppose otherwise is, in some sense, self-refuting: the claim that
there is no master-narrative is itself a master-narrative.
The enterprise is not for the faint-hearted, however. The most general
understanding of how we make sense of things—or attempt to—must en-
compass, as Moore often and correctly stresses, that very project itself. The
project is, therefore, an inherently self-referential one. Self-reference is, as
we know, and as the reference to post-modernism illustrates, wont to lead
us into contradiction. It is one aspect of this—but one central to Moore’s
book—which I wish to pursue in what follows.
2 Kant
More breaks up his book into three parts: The Early Modern Period, The
Late Modern Period I: the Analytic Tradition, and The Late Modern Period
2
II: Non-Analytic Traditions (pl., sic). I will start by looking at Moore’s
discussion of one of the philosophers from each of these parts. In this section,
we will start with Kant.
Kant explains how we can think about, and have knowledge of, phenom-
ena, that is, things of experience; for to these, we can apply concepts. But he
also holds that there are—indeed, must be—things beyond this: noumenal
objects of various kinds. Such things are required by his accounts of free will,
ethics, religion. In some sense, then, it appears that we can have thought and
knowledge about such objects. But Kant denies that we can have knowledge
of such things, since we cannot apply concepts appropriately to them. He
says (A679=B707):4
3
merely extend indefinitely. Hence, to give sense to the notion of a limit of
sense requires us to be able to give sense to what it means to be on the other
side.
Hegel puts the point perceptively:6
Thus the limit aporia; and Kant seems to fall squarely afoul of its conclusion.
Moore does his best to help Kant out of the problem. We must distinguish
between a “thick” sense of making sense, and a “thin” sense. The former
is the way in which we make sense of the phenomenal realm around us.
The latter allows us to have merely empty or regulative thoughts (EMM,
p. 153). There is certainly textual justification for this distinction. Sensing
his problem, Kant distinguished between a positive and negative notion of
noumenon—an illicit one, and a licit one whose function is merely limitative
(A255=B311). One might think of this as Moore’s thin sense.
It does not help, though. Kant was so unhappy with the distinction that
he reworked it entirely in the second edition of the Critique; but to no avail.
As Kemp Smith put it:7
Quite generally, as Moore explains, Kant requires our grasp of the noumenal
to go beyond the mere thin sense (EMM, 5.9, 5.10). In particular, it follows
6
Miller (1969), p. 134.
7
Kemp Smith (1923), p. 413f:
4
from Kant’s view that his own philosophy itself must be of such a thin kind;
but no one—Kant included—has ever understood the first Critique in this
way.
Worse: moves such as this typically simply relocate the problem of self-
referential contradiction, transferring it elsewhere.8 Grant such a thin notion
of making sense. Then there is a general notion of making sense: either
thick or thin. Kant’s own philosophical ruminations must lie somewhere
within this disjunction. But it, too, has its limits. After all, he accuses
traditional metaphysicians, the dogmatists and skeptics, of doing something
on the other side (e.g., Critique of Pure Reason, A739=B767 ff.). They
cannot, according to Kant, be doing something in the thick sense; and they
are certainly not doing what he is doing with such a thin sense; but they are
clearly doing something of which Kant can make sense. (He must understand
their enterprise in some sense, at least enough to take issue with it.) So we
are back with the Limit Argument aporia.
3 Wittgenstein
So much for Kant. Let us now turn to (the earlier) Wittgenstein. A version
of the Limit Argument is to be found explicitly in the Tractatus. In the
preface to this, Wittgenstein writes:9
[T]he aim of the book is to draw a limit to thought, or rather—not
to thought, but the expression of thought: for in order to be able
to draw a limit to thought, we should have to find both sides of
the limit thinkable (i.e. we should have to be able to think what
cannot be thought).
It will therefore only be in language that the limit can be drawn,
and what lies on the other side of the limit will simply be non-
sense.
Wittgenstein’s contrast of thought and language is a somewhat odd one, since
a thought, for Wittgenstein, simply is a proposition.10 So if the limit can
be drawn in language—albeit when one side is nonsense—it can be drawn in
thought in the same way.
8
BLoT, p. 227 ff.
9
Pears and McGuinness (1961), p. 3.
10
Tractatus 3: ‘A logical picture of facts is a thought.’
5
Anyway, the remarks presage the final dénoument of the Tractatus. States
of affairs comprise objects, constituted with a certain form. Propositions
comprise names, constituted with a certain form. Names refer to objects,
and a proposition is true if the isomorphic state of affairs exists. It follows
that only objects can be spoken of: forms, propositions, and other non-
objects cannot. Propositions show their form, the propositions they express,
and so on, and ‘what can be shown cannot be said’ (Tractatus 4.1212).
The rub, of course, is that the Tractatus says a lot of these supposedly
unsayable things. Indeed, most of its statements are of this kind. As Russell
puts it in his introduction to the English version of the Tractatus:11
6
us make maximal general sense of things. But since the sense that
it is designed to help us make is non-propositional, the means that
it uses are indirect. It works through a creative use of nonsense.
One might wonder what sort of creature it is, which is the understanding
involved in a non-propositional sense of things. Moore gives us hints that
it might be the kind of understanding that can be expressed through music
(EMM p. 240) or action (EMM, p. 255).14
One may worry here that such understanding is not essentially non-
propositional. After all, the know-how involved in riding a bike can be
expressed in terms of the laws of physics, and cognitive mechanisms which
respond to them; and the understanding of Mme Butterfly’s situation can be
described linguistically—if much more prosaically than in the music of the
last act of Puccini’s opera.
But more is at issue here. Prima facie, Wittgenstein’s views can be
expressed propositionally. They are, after all, expressed by the statements
in the Tractatus. By all the standard tests of meaningfulness, these are
completely meaningful. They are grammatical; we understand them; we can
explain them. (Philosophers teach their content to their students!) Of course,
there are things which one might call ‘hidden nonsense’: sentences such as
‘it is 04.00 at the North Pole’. Arguably such statements are not so much
meaningless as false. Chronological conventions assign no determinate time
to a point lying on all lines of longitude. A fortiori, it is not 04.00 at the
North Pole. But in any case, Wittgenstein’s statements do not seem to be of
this kind. When it is explained how geochronology works, one can see that
there is something amiss with the statement about the Pole. However, when
one grasps the ideas of the Tractatus, one’s reaction is of an entirely different
kind: the illusion—if it be an illusion—of understanding remains.
But what of the ‘very powerful arguments’—as Russell puts it—that
Wittgenstein brings in support of his position? Well, if they are indeed
nonsense, they provide absolutely no ground for supposing their conclusions
true, and so establishing that they are nonsense! Wittgenstein’s position is,
hence, self-undercutting.
14
The relation between music and the inexpressible is explored a little further in Moore
(1997), ch. 9.
7
4 Derrida
And now for something (not so) completely different: the views of Derrida
on meaning.
A fairly standard view of the way that language works is as follows.
Language talks about reality. The words of our language refer to things in
this reality (be they objects, concepts, ideas, senses); and such reference gives
language determinate meaning. (Derrida calls whatever it is that one takes
words to refer to transcendental signifieds, and the view that such signifieds
give language determinate meaning, metaphysics.)
Derrida rejects this picture entirely. Words used obtain their meaning
by relating to other words, and to other uses of the same word. Meaning
is constituted by this constantly changing flux of relations—by the play of
différance, as he puts it. There is no determinate meaning. One may interpret
words in a particular way. But interpretations are just more words, and are
themselves subject to interpretation. No interpretation has privileged status
as the correct interpretation.
There is nothing, then, that determines meaning. But when people use
language, they do mean something; and what determines that? Différance.
So there would appear to be determined (and therefore determinate) mean-
ings. Not so, says Derrida. Différance is not itself determinate; a fortiori, it
does not determine anything:15
Out of the frying pan, into the fire! Has not Derrida himself talked about
différance, in the process using that very name? And isn’t he talking about
something? Of course. He has exactly the same problem as Wittgenstein:
15
Bass (1982), p. 26.
8
he has to say what he means without meaning what he says.16 More bluntly:
Derrida has said—impossibly—what he meant.
Derrida is well aware of the situation. Here is what he says:17
I try to write (in) the space in which is posed the question of
speech and meaning: (what is) meaning to say? Therefore it is
necessary in such a space, and guided by such questions, that
writing literally mean nothing. Not that it is absurd in the way
that absurdity has always been in solidarity with metaphysical
meaning. It simply tempts itself, tenders itself, attempts to keep
itself at the point of the exhaustion of meaning.
When stripped of its Francophone frills, and expressed in blunt Anglo-Saxon,
what Derrida is saying is exactly what Wittgenstein said: his words are mean-
ingless—which gives rise to exactly the same problems as for Wittgenstein.
Again, all this is explained by Moore (ch. 20).18 Moore, however, throws
Derrida a now familiar lifeline. Derrida cannot say what he means in a ‘thick’
sense. His words, however, convey something in a ‘thin’ sense—a sense which
conveys something non-propositional (EMM, p. 537):
Why, then, if there is such a thing as non-propositional sense,
should there not be a creative if unorthodox use of language
which, given the meanings of the words in play, succeeds in con-
veying such sense? That is, why should there not be a use of
language, perhaps involving language games in what might an-
tecedently have been though of as unsuitable contexts, perhaps
involving neologisms, perhaps involving contradictions, perhaps
involving nonsense, whose effect, because of the meanings of the
words in play, is, if only as a matter of brute psychological fact,
that those who encounter it, or some of those who encounter it,
come to achieve a corresponding non-propositional understanding
of things? And if such a use of language is indeed possible, then
who is to say that much of Derrida’s work ... cannot be viewed
as a case in point?
We have been here before. If Derrida does go in for games with language, this
is unnecessary. His views can be expressed in quite ordinary and meaningful
16
As Wood (1980), p. 225, nicely puts it.
17
Bass (1981), p. 14.
18
For my own exposition of the matter, see BLoT, 14.4-9.
9
propositions—and so ones that are not nonsense. I have expressed them
thus, and so does Moore. He may use neologisms, but so do scientists,
whose meanings are (presumably) thick. Reading Derrida does indeed have
a psychological effect: we come to understand his views. Not because we
appreciate nonsense-speak; and not in the way that we might hear an odd
line of poetry, and are moved to think things by it; but because we understand
the propositions put forward. (Contradictions, we will come to in due course.)
And, again, as for Wittgenstein, the view is self-undercutting: if Derrida is
not putting forward propositions, there is just no reason to suppose that
what he says cannot be said can, indeed, not literally be said.
5 Moore Himself
So far so good. In the case of each of our three philosophers, we have seen
them saying things that, when their own views are applied to their own views,
cannot be said. Their own attempts to wriggle out of the situation seem to
be of little avail. And Moore’s rescue attempts in terms of some sort of thin
notion of making sense, don’t seem to help much either. These philosophers
are simply caught in flagrante d (el )icto, contradicting themselves.
Of course, one might say ‘so much the worse for these philosophers and
their projects’. But this is not Moore’s way. He takes it that they are
on to something. Specifically, Moore holds that if one is in the business
of making sense, there is something inevitable in the situations, something
which betokens important non-propositional sense.
There is, as far as I can see, no general argument in the book for this
conclusion, other than a sort of inductive one: all the smart philosophers who
have essayed this project have landed up in the contradictory predicament;
so that seems to be what must happen. There is just one hint of a general
argument (EMM, p. 583):
10
An answer is given in another of Moore’s books, Points of View —hereafter,
PoV.19 The argument has two stages. For the first, Moore defines transcen-
dental idealism as follows (PoV, p. 281 ff.):20
• Something is immanent iff it is in the domain of a quantifier that can
appear in one of our representations.
• Something is transcendent iff it is not immanent.
• Idealism is the view that some aspect of the form of that to which our
representations answer depends on some aspect of the representation.
• Transcendental idealism is idealism with the rider that the dependence
involved is transcendent.
One can see the views of Kant, Wittgenstein (in the Tractatus), and Derrida
as versions of transcendental idealism, so defined. The things on which rep-
resentations depend, and which are without the range of our quantifiers, are
noumena, form (etc.), and différance, respectively. And transcendental ide-
alism, so defined, is obvious self-contradictory, since it quantifies over things
over which, it says, one cannot quantify. As Moore puts it (PoV, p. 166):
Transcendental idealism ... is designed to suppress our aspirations
to make sense of what is transcendent, on the grounds that such
a thing is impossible. Yet in precisely drawing our attention to
what is transcendent, and signalling it as that whereof we should
not try to make sense, it entices us to do the very thing it is
designed to stop us from trying to do; and worse still, it requires
that we do that very thing in order to assimilate it (transcendental
idealism) in the first place.
But why accept transcendental idealism? Moore holds that we are drawn to
it in the project of making sense because it seems to explain a number of
things: ‘there is a class of problems in philosophy to which transcendental
idealism provides a general solution’ (PoV, p. 117). Specifically:21
• It can explain the necessity of certain things. (Think of Kant’s expla-
nation of synthetic a priori truths.)
19
Moore (1997). Thanks to AWM for drawing my attention to this.
20
EMM itself contains a somewhat terser definition of transcendental idealism (p. 142).
21
PoV, p. 116 ff.
11
• It reconciles various incompatible thoughts to both of which we are
drawn, such as: had sentient creatures never existed, neither would the
universe and the universe might have been such that sentience never
appeared. [The dependence on sentience is not immanent but, tran-
scendent.]22
12
can recognise good and bad arguments. Reasoning with “nonsense”, then,
has a logic. Logic presupposes contents that can be reasoned about. Say
that these are not propositions if you wish, but the cost is that the notion of
a proposition has now gone on holiday.
And to add insult to injury: if transcendental idealism is nonsense, so
is Moore’s own view. Any definition of it, or claims about what it does,
must also be nonsense. (Cf. ‘That the borogoves are mimsey has important
explanatory power.’) Nonsense, also, are the arguments marshalled in favour
of it. So they give us no reason to accept the truth of the claim that they
are nonsense.
13
Propositions such as (*) are meaningless, and simply indicate some non-
propositional sense. Here, then, we have the argument for the existence of
non-propositional sense: it provides a solution to König’s paradox.
There are important details of this argument that would need to be filled
out. Why, this time, is it that things such as (*) are meaningless? How,
exactly, do they manage to indicate something non-propositional? And what
exactly is it that they indicate? (Of course, one cannot answer the last
question by stating what it is, but one had better say something, or the
thought that they convey some sort of sense will be entirely empty.)
But set these things aside. This argument at least avoids appealing to
transcendental idealism, and in particular, the arguments for it. The ar-
guments for the inconsistency involved in Köning’s paradox are notoriously
difficult to gainsay. But the difficulties with the second stage of the argu-
ment remain. The paradoxical reasoning, for all the world, is propositional.
Anyone with a basic knowledge of set theory can follow it. There seems
to be little reason—other than that the reasoning ends in contradiction—to
suppose that it is not constituted by propositions. And, again, if the “propo-
sitions” are nonsense, they provide no ground for the claim that they are
nonsense.
Moreover, the argument shows that Moore’s position, quite generally,
faces exactly the same problem we saw Kant’s view to face at the end of Sec-
tion 2. There is, for Moore, a widest notion of making sense. Call this making
Sense (capital ‘S’). Consider the absolute infinitude of ordinals. There is no
way one can make Sense of all of them. We certainly can’t frame propositions
about most of them. And even if we can indicate—in some sense—things
about some of these, we can hardly do this about all of them. You can’t in-
dividuate the individuals of an absolutely infinite totality by actions or other
humanly available means. (A favourite theme of PoV is our finitude.) So by
the properties of ordinals, there is a least ordinal of which one cannot make
Sense. But one can makes Sense of it. We can say that it is the least ordinal
of which we cannot make Sense. This may not be a meaningful “proposition”
but, given the train of thought being pursued, it may be taken to indicate
something about the number in question. So one can make Sense of things
of which one cannot make Sense.
14
7 Contradiction
And so to contradiction. Beyond the Limits of Thought considers various
limits: the limits of expression, of iteration, of cognition, and conception.
The book argues that the self-reference inherent in the limiting constructions
makes these limits contradictory ones: there are things on the other side of
the limit which cannot be on the other. The book does not consider the limit
of what makes sense, but this limit is closely related to the limit of conception.
If one can conceive of something, one is making some kind of sense of it; and
if one can make sense of it, one must be able to conceive of it in some way
or other.27 It is not, therefore, surprising that the familiar pattern emerges
in the limit case central to Moore’s concerns. Conceiving—that is, trying to
make sense of—what is beyond this limit delivers contradiction.
Moore, as we have seen, notes the contradiction involved, but takes this
to be a sign that at least one of the statements involved in the contradiction
is meaningless. As we also saw at the end of Section 5, the thought is a hard
one to defend. Why does he not just accept the dialetheic nature of the limit
concerned?
The reason comes out most clearly in his discussion in EMM of Hegel,
who, as we saw, thought deeply about the nature of limits of this kind, and
who certainly does appear to endorse their contradictory nature.28 Moore
says the following (EMM, p. 183):
Now whenever we are confronted with a philosopher who departs
from common sense in this way, with evidently serious intent,
we seem to have a choice: either to accredit the philosopher in
question with a non-standard view of things or to accredit him
with a non-standard way of talking. I myself am attracted to the
idea that, when a philosopher goes as far as to accept a contra-
diction, then strictly speaking only the second alternative makes
sense; for unless the philosopher in question has his own idiolect,
he is violating certain basic linguistic rules ... and is not strictly
saying anything at all. ... When Hegel endorses a contradiction
27
Thus, for example, those who would argue that God is beyond conception, clearly
conceive of God in the process. See BLoT, ch. 4.)
28
BLoT, ch. 7. In general, I am very sympathetic to Moore’s exegesis of the philosophers
in his cast. But I certainly part company with him on Hegel. My understanding of Hegel,
the way his dialectic works, and the role of contradiction in it, are quite different from his.
That is too large an issue to take on here; but see Priest (1990), (2013).
15
he is violating the rules that govern the workings of some of the
words he is using: he is not using those words with their standard
meaning.
It is important to note that Moore is certainly not writing off the Hegelean
enterprise. The endorsement of contradiction is doing something—just not
one of making propositional sense. But whether that is so, is exactly now
the point at issue.
In what sense does someone who endorses contradictions violate linguistic
rules? The sentence ‘the liar sentence is both true and not true’ is a per-
fectly grammatical sentence of English, as any standard grammar will attest.
Moreover, by any such grammar, it is perfectly meaningful. According to a
standard account of conjunction and negation, it is false; but it is perfectly
meaningful. Nor do contradictions have zero content. (They say something.)
According to Frege/Russell logic, they have total content: they imply every-
thing. I might even note that it is Moore who appears to be flouting linguistic
rules, since he takes perfectly good grammatical sentences of English to be
nonsense.
Perhaps, then, the problem is not about meaning, but about endorse-
ment. Moore seems to think that someone who asserts contradictions has
ceased to speak standard English. This is just false. Native English speakers
are prepared to endorse contradictions—and be understood as doing so—as
various empirical linguistic studies show. Thus, they frequently utter contra-
dictions to indicate borderline cases.29 Maybe, one might say, such language
is metaphorical. Maybe not.30 But just like Hegel, I am a philosopher who
endorses contradictions.31 I mean these things literally, and I understand
these contradictions in exactly the same way that Moore does. I am, after
all, just as competent a native english-speaker as he. And if he rejects di-
aletheism, it is because he understands exactly what I say: it is in virtue of
this that he takes me to be mistaken.
Indeed, Moore takes such sentences not only to be false, but to be false
in virtue of what the words ‘and’ and ‘not’ mean. I do not. That is a
disagreement. It does not necessarily mean that he and I have different un-
derstandings of the meanings of those words. I do not know his views on the
matter, but I assume that he would endorse some version of the standard
29
See Ripley (2013).
30
See Priest (2010).
31
See, e.g., Priest (1987).
16
truth-conditional account of the meanings of these connectives. So do I. Now,
in the semantics of paraconsistent logics like LP (and First Degree Entail-
ment), the truth and falsity conditions—and so meaning—of the connectives
are exactly the same as in Frege/Russell logic.32 Where Frege/Russell logic
and LP part company is simply about whether truth and falsity can overlap
(or underlap). This is not a semantic disagreement, but a metaphysical one.
Compare Aristotle’s view that contingent statements about the future are
not yet true or false (so that truth and falsity underlap). This is a metaphys-
ical view about the determinacy of the future. And one may well take the
fact that there are apparently obviously correct principles, such as the T -
schema, which seem to deliver contradiction, to show that truth and falsity
do overlap.
And even if Moore and I do have different views about the meanings
of the connectives, the disagreement is hardly one where he is obviously
right. Developments in the philosophy of language and logic in the last 40
years have taught us that delivering an account of meaning—even of just
the connectives—is a highly fraught enterprise. It is not even clear what the
right form of such an account should be. (Nor would the fact that Moore
and I have different theories of meaning for the connectives imply that we
mean the words differently: we just have different theories about what that
meaning is.) However, all this is clearly far too big an issue pursue here.
So I shall merely close this part of the discussion by noting that whatever
disagreement there is here between Moore and myself cannot be brushed
aside with a vague appeal to “linguistic rules”.
8 And in Conclusion...
It may seem that Moore and I are a million miles apart; and in some sense,
we are. In another sense, the difference between us is a very small one.
We both agree that accounts of making sense, when applied to themselves,
appear to generate contradictions at the limits of sense-making. We also
agree that this is no mere philosophical frill, but at the very core of a central
philosophical enterprise. The difference between us is that where he sees
nonsense attempting to express the ineffable, I see a plain dialetheia.
Both paths take contemporary philosophy off in an unorthodox direc-
tion. (What interesting philosophy does not?) But Moore’s approach seems
32
See, e.g., Priest (1998).
17
fraught with the problems that I have rehearsed. The arguments to the ef-
fect that the sentences in question are meaningless are self-undercutting; and
there is no independent argument for this claim: they can appear meaningless
only to someone already in the grip of a philosophical theory. Moreover, as
we saw in Section 6 with respect to König’s Paradox, Moore’s approach seems
to generate the same sort of “revenge” problem which he takes himself to be
avoiding. A dialetheic solution, on the other hand, simply requires one to ac-
cept that some contradictions are true. That this is radical in contemporary
Western philosophy can hardly be denied. But the Law of Non-Contradiction
was never as rationally grounded as those who like to trumpet it suppose.
As far as I can see, the main thing that stands in the way of the dialetheic
understanding of the situation is this last bit of Aristotelian dogma.33
So:—at the limit of making sense, we stop making sense; but we can make
sense of some of the things beyond the limit: we can conceive of that of which
we cannot conceive. Yes—and we can make sense of that too.
References
[1] Bass, A. (trans.) (1981), Positions, Chicago, IL: University of Chicago
Press.
18
[8] Pears, D. F., and McGuiness, B. F (transs.) (1961), Tractatus-Logico
Philosophicus, London: Routledge and Kegan Paul.
[10] Priest, G. (1990), ‘Dialectic and Dialetheic’, Science and Society, 53:
388-415.
19