Blackwell Publishing Ltd.Oxford, UK and Malden, USAANALAnalysis0003-26382005 Blackwell Publishing Ltd.

April 200565216773ArticlesBradley Armour-garb & Graham Priest

Analetheism: a Pyrrhic victory

Analetheism: a Pyrrhic victory

Bradley Armour-Garb & Graham Priest

1. Introduction
In a recent article Jc Beall and David Ripley (2004), bending a term
of Terry Parsons’s (1990), describe a position concerning the paradoxes
of self-reference that they call ‘analetheism’, and claim that it would
appear to be just as good as the dialetheic account advocated by Priest
(1987) – hereafter IC. They say:
As far as we can see, the analetheist achieves whatever expressive
virtues that the dialetheist achieves; and she also partakes of the same
sort of expressive vices as the dialetheist. What could tip the scales
in favour of one position over the other? We do not know. (33)
The present note provides an answer.
IC provides what is, in effect, a three-valued logic in which the middle
value is interpreted as both true and false. And since it is at least true, it
is designated. Analetheism adopts exactly the same logic, but interprets
the middle value as neither true nor false. What distinguishes analetheism
from more simple-minded truth-value gap theories is precisely that it takes
the middle value to be designated. In particular, things with the middle
value are just as assertable as things that are true.1

2. An expressive difficulty
According to dialetheism, paradoxical sentences are both true and false
(i.e. have a true negation). Indeed, some sentences, such as the liar paradox
in the form l: ÿT·lÒ, are both true and not true: T·lÒ& ÿT·lÒ.2 It follows
(by De Morgan laws, which are valid in the logic) that ÿ(T·lÒ ⁄ ÿT·lÒ).
Now, Beall and Ripley say (32):
The intended interpretation of ‘T·lÒ’ (that l is true) renders ‘ÿ(T·lÒ
⁄ ÿT·lÒ)’ as the claim that l is neither true nor not true, certainly
not something the dialetheist wants to say given her aversion to truth-
value gaps.

1
‘Given that some sentences of the form A&ÿA may be designated (and hence
assertable), ...’ (Beall and Ripley 2004: 32).
2
We use T for the truth predicate, and angle brackets as a name-forming device. Beall
and Ripley use l as a name, not a sentence, and so write T l.

Analysis 65.2, April 2005, pp. 167–73. © Bradley Armour-Garb and Graham Priest
168 bradley armour-garb & graham priest

Now, it is not clear why the dialetheist should be averse to the claim in
question. There are no truth-value gaps in the dialetheic semantics; but
neither does this sentence say that l is a gap. This would be expressed by
the sentence ÿ(T·lÒ ⁄ ÿT·lÒ), which is not forthcoming without the
principle that T·lÒ Æ ÿT·lÒ, which IC (4.9) rejects.
So why should it be thought so? The sentence is, after all, true: both
disjuncts are false; so, therefore, is their disjunction. The problem, if there
is one, is in thinking that the sentence means something that it does not,
namely that it rules out l’s being true (and not true). In consistent con-
texts, it does so, but things are not always as they seem in inconsistent
contexts. In a paraconsistent logic, ÿa does not rule out a. In particular,
ÿ(T·lÒ ⁄ ÿT·lÒ) does not rule out T·lÒ.
Any residual discomfort one might experience here can be accounted
for by pointing out some facts about conversational implicature. When
we assert, we are bound by the Gricean conversational maxim to say
everything that is relevant. Thus, simply to assert ÿa, when a and ÿa are
both true, is misleading. A hearer may reasonably draw the conversational
implicature that a is simply false. Better to say a & ÿa. To assert just
ÿ(T·lÒ ⁄ ÿT·lÒ) is misleading in a similar way, since T·lÒ ⁄ ÿT·lÒ is also
true. Better to say ÿ(T·lÒ ⁄ ÿT·lÒ) & (T·lÒ ⁄ ÿT·lÒ).
As Beall and Ripley note, if dialetheism has any problem here, analethe-
ism has a similar problem. An analetheist asserts that ÿ(T·lÒ ⁄ ÿT·lÒ).
By De Morgan’s laws, this is equivalent to T·lÒ & ÿT·lÒ, which appears
to say that l both is and is not true. But the reply is the same: even though
we are dealing with truth-value gaps, contradictions are still acceptable.
In particular, T·lÒ & ÿT·lÒ does not rule out ÿ(T·lÒ v ÿT·lÒ). Similar
points about conversational implicature also apply.

3. The T-schema
Let us now turn to reasons for preferring dialetheism to analetheism. IC
subscribes to the T-schema:

• T·aÒ ´ a

It does not, however, accept its contraposed form (IC, 4.9):

• ÿT·aÒ ´ ÿa
That ÿa does not entail that a is not true: it might be a truth-value glut.
Playing the other side of the street, analetheism accepts the contraposed
form of the T-schema, but not the T-schema itself. That a does not entail
that a is true: it might be a truth-value gap.
It is, in fact, open to both parties to accept both the T-schema and its
contraposed form, as Beall and Ripley themselves point out at the end of
 analetheism: a pyrrhic victory 169

their paper. We will return to this later. For the moment let us just consider
analetheism in the form in which they present it. An evident weakness of
this is just that it does not endorse the T-schema, the ability to do which
is one of the great advantages of dialetheism.
It might be contested that dialetheism suffers from an analogous weak-
ness: it does not accept the contraposed form of the schema. But this
principle is not in the same league. From Aristotle, through Tarski, to
contemporary deflationism about truth, it is the T-schema that is taken
as fundamental to truth, not its contraposed form. To say of what is that
it is, this is the truth.3 The contraposed form does not even say what truth
is; it just comes along for the ride – on the back of contraposition, if that
holds for the biconditional employed in the schema.
This is one weighty consideration.

4. Truth and assertion

We now come to a weightier one: the evident oddity of designating the
middle value if this is interpreted as neither true nor false. In a many-
valued logic designation picks out those sentences that are acceptable or
assertable (given the interpretation in question). More generally, validity
is defined in terms of designation-preservation,4 validity being important
because it preserves the property of being a sentence that we are to accept.
Now, it is a widely held view that truth is the telos of acceptance and
assertion. That is, truth just is what correct assertion aims at. If this is
right, designated values must be truths of some kind. The view that truth
is a telos of assertion is not only one that appears to recommend itself to
common sense; it has been widely defended, e.g. by Dummett (1959);
Brandom (1983); Priest (IC, 4.6); Williamson (1996).5 It is not our aim
to reproduce the various arguments employed by these writers. The point
we wish to make here is simply that Beall and Ripley cannot subscribe to
this view. They are certainly under no illusion about this. They point out
that on the account in question one should assert things that are neither
true nor false, and so not true. But they do not attempt to contest the
arguments concerning truth as the telos of assertion.
The oddity of the claim that one should assert things that are neither
true nor false is also fairly evident. It is not just certain self-referential
sentences that might be thought to be neither true nor false. There are a
whole panoply of sentences that have been claimed to have this status:
3
Indeed, Beall himself is a deflationist, and endorses the T-schema. See, e.g., Beall
(2004).
4
See, e.g., Priest 2001: 7.2.
5
In particular, then, it is not ‘dogmatically’ that we accept the view that it is truths
that are to be asserted, contra what Beall and Ripley say (34).
170 bradley armour-garb & graham priest

future contingents (‘there will be a sea battle tomorrow’), denotation

failures (‘the King of France is bald’), nonsense (‘Twas brillig and the
slithey toves did gyre and gimble in the wabe’), and so on. It follows from
the analetheist norm that these should be asserted too – which would seem
quite wrong. Of course, one may deny that these sentences are neither
true nor false. Maybe they are one and all false. But the analetheist is still
committed to the conditional that if any of these examples did turn out
to be neither true nor false, we would have to assert them. The counter-
intuitiveness of this is plain enough.

5. Falsity and Assertion

Beall and Ripley might suggest6 that dialetheism also has some counter-
intuitive consequences concerning norms of assertion. Specifically,
dialetheists cannot subscribe to the norm of assertion to the effect that
one should refrain from asserting what is false, since they are prepared to
assert something that is false, as long as it is also true. An analetheist, by
contrast, does respect this norm.
Care must be exercised at this point, however. The notion of falsity may
reasonably be taken in two ways. To say that a is false might be to say
that it has a true negation (which is the way we have interpreted it till
now); or it might be to say that a is not true. So the norm in question
might be:
• one should refrain from asserting what has a true negation
or it might be:
• one should refrain from asserting what is not true
Of course, given classical logic the two notions of falsity are equivalent.
It is therefore easy to confuse these two norms. But once gaps and gluts
are on the scene, the two notions of falsity are quite distinct. One must
therefore ask which of these two norms it is that is at issue here.
It seems fairy clear that it should be the second of these. Indeed, this is
just the flipside of the truth-norm. The true is just what we should assert;
the untrue is, then, just what we should not. If the untrue and the false
are the same, then the first-falsity norm is fine. But if one takes it that
there are truth-value gluts, then the norm will appear too strong. The fact
that something has a true negation does not rule out its being true, and
so its being assertable. And if one takes it that there are truth-value gaps
(at least in the usual way), then the norm will appear too weak: there will
be things one ought to refrain from asserting even though they do not
have a true negation (gappy sentences).
6
As they did in correspondence.
 analetheism: a pyrrhic victory 171

The first falsity-norm may get its pull from two (mistaken) sources. The
first is from the classical confusion of what is not true with what has a
true negation. The second is from the considerations concerning conver-
sational implicature that we mentioned in §2. If we assert that a has a
true negation, that is, ÿT·aÒ, and so, by implication, ÿa, we may well
conversationally infer that a is true-only, that is that ÿT·aÒ. In this case,
we can then apply the first falsity-norm to infer that one should refrain
from asserting a. But this rests on a conversational implicature, not an
entailment.
Now that we have this straight, we can ask whether a dialetheist does
abandon the falsity norm. The answer is no. A dialetheist (but not an
analetheist) can well subscribe to the claim that if something is untrue it
ought not to be asserted. They can also, as we have just noted, subscribe
to the first-falsity norm, but only as a conversational implicature. It is the
second-falsity norm that is unreservedly correct.
It is worth noting that the second falsity-norm may give rise to dilemmas
on occasions. As we have already observed, paradoxes of self-reference
are true and false (have a true negation); but a few – such as the liar
paradox as formulated in §2 – are stronger: they are true and not true.
For such a paradox, we both ought to assert it, since it is true, and ought
not to assert it, since it is not true. Well, life is like that sometimes: we do
face dilemmas where you are damned if you do and damned if you don’t.7
Actually, dilemmas of this kind seem to occur independent of the falsity-
norm. Consider the sentence ‘It is not the case that this sentence is
assertable’, i.e.:
• a: ÿ·aÒ is assertable.
Suppose that a is false; then a is assertable. But it can’t be assertable if it
is false. So it must be true (assuming the Law of Excluded Middle, which
both the dialetheism and the analetheist endorse). Thus, it is not assert-
able. But we have just established that it is true, so it is. You ought and
ought not to assert a.8
At any rate, the important point in the present context is that a dialethe-
ist can subscribe to the falsity-norm, correctly understood.

6. A Pyrrhic victory
We have now seen that there are a number of reasons to prefer dialetheism
to analetheism. But let us end with a final observation. As Beall and Ripley
7
The possibility of dilemmas concerning legal and moral norms is discussed in IC,
ch. 13. The dilemma concerning the falsity-norm is discussed further in Priest 1993.
8
This is a version of the ‘irrationalist paradox’. For this and other paradoxes of
rationality that have nothing to do with self-reference at all: see Priest 2002.
172 bradley armour-garb & graham priest

note, dialetheism and analetheism are exactly the same, except that where
one says ‘both true and false’ the other says ‘neither’. Is this anything more
than a nominal difference? If the truth-norm obtains, then clearly not.
Since the middle value is designated, things that have this value are true
(as are their negations). So what Beall and Ripley call ‘neither true nor
false’ is simply a species of truth.
If the norm of truth does not hold, this may not be the case. However,
it remains true that analetheists are still prepared to go around asserting
contradictions in the full knowledge that they are contradictions, and with
no caveats. This gives dialetheists pretty much everything they ask for.
Call such things neither true nor false if you like; it doesn’t really matter.
Indeed, it is not really clear what the rationale for, or virtue of, appealing
to truth-value gaps is if you end up asserting contradictions anyway.
(Normally, the whole point of appealing to the notion, in the context of
the paradoxes of self-reference, is to avoid inconsistency.)
The point, we think, is pretty much conceded by Beall and Ripley at
the end of their paper. They point out that an analetheist can, in fact,
subscribe to both the T-schema and its contraposed form (as can the
dialetheist), making a and T·aÒ generally inter-substitutable. And if one
does this there are really no differences of any significance left between
the two views; they are, as Beall and Ripley themselves say, ‘no longer
distinguishable’ (34).
Indeed so. If analetheism has any kind of victory over dialetheism it is
an entirely Pyrrhic one.9

University at Albany, SUNY

Albany, New York 12222, USA
[email protected]

University of Melbourne
Melbourne, Vic. 3010, Australia

and

University of St Andrews
St Andrews KY16 9AL, UK
[email protected]

9
Thanks go to Jc Beall and Dave Ripley for their comments on earlier drafts of the
paper.
 analetheism: a pyrrhic victory 173

References
Beall, Jc 2004. True and false – as if. Ch. 12 of G. Priest, Jc Beall and B. Armour-
Garb (eds.), The Law of Non-Contradiction: New Philosophical Essays, Oxford:
Oxford University Press.
Beall, Jc and D. Ripley. 2004. Analetheism and dialetheism. Analysis 64: 30–35.
Brandom, R. 1983. Asserting. Nôus 17: 637–50.
Dummett, M. 1959. Truth. Proceedings of the Aristotelian Society 59: 141–62; repr.
as ch. 1 of M. Dummett, Truth and Other Enigmas. London: Duckworth.
Parsons, T. 1990. True contradictions. Canadian Journal of Philosophy 20: 335–54.
Priest, G. 1987. In Contradiction: a Study of the Transconsistent. The Hague: Martinus
Nijhoff. Second (extended) edition, Oxford: Oxford University Press, forthcoming.
Priest, G. 1993. Can contradictions be true? Proceedings of the Aristotelian Society,
Suppl. Vol. 67: 35–54.
Priest, G. 2001. An Introduction to Non-Classical Logic. Cambridge: Cambridge
University Press.
Priest, G. 2002. Rational dilemmas. Analysis 62: 11–16.
Williamson, T. 1996. Knowing and asserting. Philosophical Review 105: 489–523.