V

C 2018 The Philosophical Forum, Inc.

WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF

NON-CONTRADICTION AGAINST ITS CONTRARY?
DANIEL COREN

I. INTRODUCTION

Consider the following argument:

Premise 1: All human beings are animals.
Premise 2: Not all human beings are animals.
Conclusion: Therefore Socrates is an animal.

The argument above is unsound. But according to modern propositional logic,

the argument is valid.1 There is not a semantic interpretation in which the argu-
ment’s premises are both true and its conclusion is false. More specifically, if P1
is true then P2 is false, and vice versa. As a result of the fact that P1 and P2 can-
not both be true, any well-formed formula—or combination of well-formed for-
mulas—could take the place of the sentence currently in use as the conclusion
without making the argument invalid. “Socrates is a chess master” could be used
as the conclusion, and still validity would be preserved, as could, “The square
root of forty-nine is eighteen” and/or, “You’re both alive and not alive.” For a
more general illustration of the point, consider a proof of the claim that from the
1
By “modern propositional logic” or “modern logic” I mean classical first-order logic, also known as
“classical elementary logic.” See Shapiro (S. Shapiro, “Classical Logic,” The Stanford Encyclopedia
of Philosophy (2013) http://plato.stanford.edu/entries/logic-classical/.) for a nice description and
explanation of the basics of classical first-order logic. With u, :u taken as contradictory opposites,
and w taken as any formula at all, Shapiro describes the principle of explosion as follows: “some
logicians introduce a rule to codify a similar inference: If C1 ‘ u and C2 ‘ :u, then for any formula
w, C1, C2 ‘ w.”

39
 DANIEL COREN

truth of a contradiction—a conjunction consisting in a sentence and its nega-

tion—the truth of any other sentence follows:
(1) u Ù :u Premise

(2) u 1, conjunction elimination

(3) u v w 2, disjunction introduction
(4) :u 1, conjunction elimination
(5) w 3,4, disjunction elimination

The proof above relies on three basic rules of inference and proves: {u Ù :u} !
w.2 It’s entrenched in modern propositional logic as a principle called ex falso
quodlibet or, more colloquially, explosion. With explosivity, modern classical
logic has a simple answer if one asks, “Can I accept just one or just some contra-
dictions?”3 The answer is: “No. If a single contradiction is true, then all contra-
dictions are true.”4
But Aristotle cannot accept explosivity. In his defense of the principle of non-
contradiction (PNC) in his Metaphysics C.4, Aristotle argues that if some rather than
o). He
all contradictions are true, then the exceptions will be agreed upon (homologe^
then dismisses the possibility without further discussion (C.4 1008a8–11). Instead,
as was noted first in Lukasiewicz5 and subsequently in Castagnoli,6 Dancy,7

2
This proof owes its origin to C.I. Lewis (C. I. Lewis, “Alternative Systems of Logic,” The Monist 42
(1932): 481–507. Reprinted in Lewis (1970) 400–19).
3
The position that there are some true contradictions has in recent decades been known as dialetheism,
and is motivated at length in Priest (G. Priest, Contradiction (Oxford: Oxford UP, 2006b)).
4
So blocking explosion requires rejecting at least one of conjunction elimination, disjunction elimina-
tion, and disjunction introduction as inference rules. This is the approach taken by some paraconsis-
tent logics such as dialetheism. Dialetheism holds that there are some true contradictions, such as
those generated by liar sentences. But dialetheism also maintains that not all contradictions are true.
So dialetheists are not necessarily also trivialists, though trivialists are necessarily also dialetheists.
For trivialism says everything is true, and dialetheism falls under the scope of everything. Prominent
contemporary logicians including Graham Priest and J. C. Beall, for example, block explosion and
avoid trivialism by rejecting disjunction elimination. They block, in other words, entailment from
the position held by Aristotle’s second-most radical opponents (“Some but not all contradictions are
true”) to the position held by Aristotle’s most radical opponents (“Every statement is true (and
false”). They hold the former but not the latter.
5
J. Lukasiewicz, (in Polish) “On the Principle of Contradiction in Aristotle,” Review of Metaphysics
24 (1910): 485–509; “Aristotle on the Law of Contradiction,” Articles on Aristotle, vol. 3,
Metaphysics, ed. Barnes, Schofield, and Sorabji (London: Duckworth 1979) 50–62.
6
L. Castagnoli, Ancient Self-Refutation (Cambridge: Cambridge UP, 2010).
7
R. M. Dancy, Sense and Contradiction: A Study in Aristotle (Dordrecht, the Netherlands: D. Reidel,
1975).

40
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

Gottlieb,8 Kirwan,9 Politis,10 Priest,11 Rapp,12 and Wedin,13 Aristotle’s defense of

PNC seems almost exclusively devoted to the question “Can all contradictions be
true?” as opposed to the more salient question “Can even one contradiction be true?”
But just because not all contradictions are true, so that PNC’s contrary is false, it is
not necessarily the case that all contradictions are false. PNC’s contradictory might
still be true. So PNC might still be false. That not all basketball players are short
does not entail that all basketball players are not short. Aristotle does not seem to
succeed, then, in defending PNC.
I argue that the criticism above is not fatal to Aristotle’s defense of PNC.
Section II presents and briefly analyzes Aristotle’s different formulations of PNC
in order to get clearer about what, precisely, he is defending. Section III explains
in more detail why Aristotle’s defense of PNC appears to fail. Section IV shows
that Aristotle is keenly aware of the distinction between contraries and contradicto-
ries, and that he distinguishes between PNC’s contradictory and PNC’s contrary in
C.4. The paper then gives three main reasons to think that Aristotle’s defense of
PNC is not crippled by its focus on PNC’s contrary. First, in Section V, I argue
that one vital reason Aristotle attacks PNC’s contrary is that he took his real oppo-
nents such as Anaxagoras, Cratylus, Democritus, Empedocles, Heraclitus,
Parmenides, Protagoras, Xenophanes, and even Homer to hold views deeply con-
nected with PNC’s contrary. Second, in Section VI, I explain why Aristotle’s
defense of PNC must be a particular kind of indirect defense than a direct demon-
stration. Third, in Section VII, I propose a reading on which Aristotle gives a con-
vincing defense of PNC by demonstrating the reliance of coherent communication
on non-contradiction. More generally, then, I aim to go some way towards
motivating a reading of Aristotle’s defense of PNC that is not only exegetically
and historically plausible but also philosophically attractive.

II. WHAT IS ARISTOTLE’S PNC?

The following are four formulations of PNC Aristotle gives in Metaphysics

C.3–6:
8
P. Gottlieb, “The Principle of Non-Contradiction and Protagoras: The Strategy of Aristotle’s
Metaphysics IV 4,” Proceedings of the Boston Area Colloquium in Ancient Philosophy VIII (1994):
183–209; Gottlieb, (P. Gottlieb, “Aristotle on Non-contradiction,” The Stanford Encyclopedia of
Philosophy (2015) http://plato.stanford.edu/entries/aristotle-noncontradiction/.)
9
C. Kirwan, Aristotle; Metaphysics, Books C, D, E, 2nd ed. (Oxford: Oxford UP, 1993).
10
V. Politis, Aristotle and the Metaphysics (London: Routledge, 2004).
11
G. Priest, Doubt Truth to Be a Liar (Oxford: Oxford UP, 2006a).
12
C. Rapp, “Aristoteles € uber die Rechtfertigung des Satzes vom Widerspruch,” Zeitschrift f€
ur
philosophische Forschung 47 (1993): 521–41.
13
M. V. Wedin, “A Curious Turn in Metaphysics Gamma: Protagoras and Strong Denial of the
Principle of Non-Contradiction,” Archiv fur Geschichte der Philosophie 85 (2003): 107–30.

41
 DANIEL COREN

[A]14 “Evidently then such a principle is the most certain of all; which principle this is, we proceed
to say. It is, that the same attribute cannot at the same time belong and not belong to the same
subject in the same respect; we must presuppose, in face of dialectical objections, any further
qualifications which might be added” (100517-22).15
[B] “It is impossible that contrary attributes should belong at the same time to the same subject (the
usual qualifications must be presupposed in this proposition too)” (1005b26–28).16
[C] “It is impossible for anything at the same time to be and not to be” (1006a3–4).17
[D] “The most indisputable of all beliefs is that contradictory statements are not at the same time
true” (1011b13–14).18

[D] expresses PNC with respect to the superlative indisputability of a belief

(doxa), whereas [A] through [C] do not mention a belief or anything closely
related. Rather, [A] through [C] express PNC with respect to the impossibility of
the simultaneous truth of contradictory statements, or the impossibility of the
same thing being both said and not said of the same thing (to auto). Any particu-
lar belief related to PNC, let alone the indisputability of such a belief, is absent
from [A] through [C]. Consider a world at which [A] through [C] are true, but
[D] is false. So it is true at this world that “It is impossible for anything at the
same time to be and not to be.” It’s also true at this world that contradictory state-
ments are false, and that to say that something, x, can be both F and not-F, is
false. But it’s also true at this world that the belief that contradictory statements
cannot be simultaneously true is not the most indisputable belief. Rather, at this
world the belief in the following is just as indisputable as the belief in PNC: for
any two objects such as a sound and a color, “each of them is different from the
other and the same as itself [. . .] and that both together are two, and each of them
is one” (Theaetetus, 185a).
The world described above does not seem terribly implausible: it might well
be argued at the actual world that the belief in that principle is just as indisput-
able as the belief in PNC. Indeed, such a world might be close to Aristotle’s. For

14
All Greek text from the Metaphysics is taken from W. D. Ross, Aristotle’s Metaphysics, 2 vols.
1924, (reprinted 1970 [of 1953 corr. edn.]).
15
ὅsi lὲm oὖm bEbaios ash ἡ soiaύsh parῶm, dῆkom: sί1 d᾽ ἔrsim aὕsh, lEsὰ saῦsa kegxlEm. sὸ
gὰq aὐsὸ ἅla ὑp aqvEim sE jaὶ lὴ ὑp aqvEim ἀdύmasom sῷ aὐsῷ jaὶ jasὰ sὸ aὐsό (jaὶ ὅra
ἄkka pqordioqiraίlEu᾽ ἄm, ἔrsx pqordixqirlema pqὸ1 sὰ1 kogijὰ1 dtrvEqEίa1) (1005b17–
22). Kirwan refrains from the use of “attribute” in his translation of 1005b19–22: “For the same
thing to hold good and not to hold good simultaneously of the same thing and in the same respect is
impossible (given any further specifications which might be added against the dialectical
difficulties)” (Kirwan 1993: 7).
16
Eἰ dὲ lὴ ἐmdevEsai ἅla ὑp aqvEim sῷ aὐsῷ sἀmamsίa (pqordixqίrux d᾽ ἡlῖm jaὶ saύsῃ sῇ
pqos arEi sὰ Eἰxuόsa).
17
ἡlEῖ1 dὲ mῦm EἰkήualEm ὡ1 ἀdtm asot ὄmso1 ἅla Eἶmai jaὶ lὴ Eἶmai.
18
ὅsi lὲm oὖm bEbaios ash dόna parῶm sὸ lὴ Eἶmai ἀkhuEῖ1 ἅla sὰ1 ἀmsijEilema1 u arEi1.

42
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

Aristotle seems to implicitly rely on the principle of identity throughout his

defense of PNC. I think it’s fairly clear, then, that it’s possible and perhaps even
plausible that [D] might be false even if [A] through [C] are true.19 Due to the
strength of the claim in [D] about the belief peculiar to its formulation, one might
argue that while [D] is connected with [A] through [C], [D] is a different princi-
ple altogether. There are no doubt other interesting differences among [A]
through [C]. Much more could be said here. But since my focus lies elsewhere,
I’ll not part with convention: I’ll take [A] as Aristotle’s representative formula-
tion. It can be stated a little more precisely as follows. For any object, x, and any
predicate, F, it’s not the case that x is both F and not-F: (8x)(8F)(Fx & Fx).

III. WHY ARISTOTLE’S DEFENSE OF PNC SEEMS TO FAIL

Given that PNC is such a strong claim, namely, (8x)(8F)(Fx & Fx), we
would expect Aristotle to conclude its defense along the lines of, “It is clear from
what has been said that for all things and all that can be predicated of things,
nothing can in the same respect be predicated and not predicated of the same
thing.” Instead, it is with a remarkably weak claim that Aristotle concludes his
string of arguments in Metaphysics C.4 in defense of PNC: “We shall have got
rid of the unqualified doctrine which would prevent us from determining any-
thing in our thought” (1009a3–4). Nowhere does Aristotle explain why his oppo-
nents must accept that not just some but all contradictions are false. Rather, we
find Aristotle arguing, even as late as C.6, for claims as weak as: “if not all things
are relative, but some exist in their own right (jau᾽ aὑsά), not everything that
appears will be true” (1011a18–19).
Lukasiewicz20 appears to be have been the first to note that Aristotle’s defense
of PNC attacks PNC’s contrary rather than PNC’s contradictory: “Aristotle
proves not that the mere denial of the principle of contradiction would lead to
absurd consequences, rather he attempts to establish the impossibility of the
assumption that everything is contradictory.” Over a century later, Priest21 argues
19
It is perhaps worth noting, to help explain the difference between [D] and the other formulations,
that [D] occurs in C.6, whereas [A] through [C] occur much earlier on in Aristotle’s discussion of
PNC at C.3 and C.4. By the time we reach [D] in C.6 the discussion has explicitly shifted to an attack
on the radical Protagorean opponents who hold that everything in the world is just as it appears to
everyone and anyone. Besides my own discussion in section V of these more radical opponents, see
Lee (M. Lee, Epistemology after Protagoras: Responses to Relativism in Plato, Aristotle, and
Democritus (Oxford: Oxford UP, 2005), especially chapters 4, 6, and 7, for a detailed exegesis,
examination and discussion of Protagorean theories and the responses from Plato, Aristotle (in C.4
through C.6), and other figures such as Democritus.
20
Lukasiewicz (1910).
21
Priest (G. Priest, “Dialetheism,” The Stanford Encyclopedia of Philosophy (2013) http://plato.stan-
ford.edu/entries/dialetheism/.)

43
 DANIEL COREN

that Aristotle’s defense of PNC “cheerfully slides between attacking the claim
that some contradictions are true [the contradictory of PNC] and the position that
all contradictions are [the contrary of PNC].”22 Gottlieb23 puts the problem as
follows:
In arguing for PNC we should expect Aristotle to be addressing someone who thinks that there is at
least one case in which it is possible for the same thing to belong and not to belong to the same thing
in the same respect. Instead, Aristotle spends most of [Metaphysics C.4] addressing someone who
thinks that in every case (i.e., take any subject and property you like) it’s possible for the same thing
to belong and not to belong to the same thing in the same respect. Aristotle appears to be arguing
against someone who purports to believe the contrary and not the contradictory of PNC.

Given that Aristotle’s PNC is such a strong claim, namely, (8x)(8F)(Fx &
Fx), even a single counterexample is enough to falsify PNC. For instance, out
of all the trillions upon trillions of objects and predicates in the entire universe, if
there is nothing that falsifies PNC except just one snowflake that is both symmet-
rical and not symmetrical in the same respect, then PNC is false. In short, the
position that just some contradictions are true—hereafter referred to as SOME
and formalized as, (9x)(9F)(Fx & Fx)—or even that just one contradiction is
true, falsifies PNC.24 But Aristotle appears to devote his defense of PNC to the
position that all contradictions are true—hereafter referred to as ALL and formal-
ised as, (8x)(8F)(Fx & Fx)—rather than SOME.25 At the very least, he does not
demarcate or explain where and how he is attacking SOME rather than ALL.
But it seems that Aristotle must clearly demarcate and explain where and how
he is attacking SOME rather than ALL. For even if ALL is false, SOME might still
be true. The fact that not all contradictions are true does not entail that all contra-
dictions are false. So even if Aristotle has shown that ALL is false, he has not
thereby shown that SOME is false. So PNC might still be false. Modern classical
logic has explosivity, as was shown in Section I, so that SOME entails ALL. That
22
The problem is also noted in some detail in Kirwan (1993: 101–4). While Castagnoli (2010) is not
specifically interested in the slide, he claims that “starting from 1007b18 the focus [of Aristotle’s
defense of PNC] shifts to the extreme thesis that ‘contradictories are all simultaneously true of the
same thing’. . .” (Castagnoli 2010: 68). Politis (2004: 138–41; 153–55) also comments on the slide.
Priest (2006a) comments on it in some detail. Wedin (2003) notes and discusses the problem. Wedin
(M. V. Wedin, “On the Use and Abuse of Non Contradiction: Aristotle’s Critique of Protagoras and
Heraclitus in Metaphysics Gamma 5,” Oxford Studies in Ancient Philosophy 26 (2004b): 213–39)
also mentions the problem, though only in passing.
23
Gottlieb (1994).
24
As Kirwan (1993: 104) notes, it’s not clear that Aristotle would have distinguished between (9x)
(9F)䉫(Fx & Fx) and (9x)(9F)(Fx & Fx), or between (8x)(8F)w(Fx & Fx) and (8x)(8F)(Fx &
Fx).
25
Similar observations can be found in Castagnoli (2010: 68–9), Cohen (S. Marc Cohen, “Aristotle on
the Principle of Non-Contradiction,” Canadian Journal of Philosophy 3 (1986): 359–70), Kirwan
(1993: 102–3), Lee (2005: chapter 6), Politis (2004: 139–41, 153–5), and Priest (2006a: 11–12).

44
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

makes a refutation of ALL sufficient. But Aristotle cannot accept explosivity.26

Also, he gives clear evidence at C.4 1008a8–11—as will be discussed pres-
ently—that he considers SOME an independent possibility. So Aristotle’s
defense must show that SOME is false. But it does not appear to do so.

IV. WHY ARISTOTLE’S APPARENTLY EXCLUSIVE FOCUS ON ALL IS

NOT A BASIC, LOGICAL ERROR

During his defense of PNC, Aristotle makes a clear distinction between SOME
and ALL:
Again, either the theory is true in all cases, and a thing is both white and not-white, and existent and
non-existent, and all other assertions and negations are similarly compatible [5df ALL], or the the-
ory is true of some statements and not of others [5df SOME]. And if not all, the exceptions will be
agreed upon; but if of all [. . .] (1008a8–11).27

After making this distinction, Aristotle dismisses SOME with great ease. All he
says about SOME, other than its definition, is that it’ll follow as a consequence
that there will be certain agreed-upon points of truth. Kirwan28 takes an unchari-
table view of this dismissal, claiming that Aristotle “dismisses [SOME] with the
inadequate comment that it narrows the field of dispute.” Kirwan says no more

26
Curiously, Aristotle appears to describe and perhaps endorse explosivity in the Physics in reference
to “refuting the merely contentious” arguments of Melissus and Parmenides, whose “premises are
false and conclusions do not follow [. . .] Or rather the argument of Melissus is gross and offers no
difficulty at all: accept one ridiculous [ἄsopo1] proposition and the rest follows – a simple enough
proceeding.” (Phys.185a8–12). Aristotle seems to be arguing that because Melissus accepts one ato-
pos or absurdity, refuting Melissus is simple enough, since such a refutation involves merely point-
ing out that all the other atopoi follow from the acceptance of that first atopos. But it’s unclear to me
whether Aristotle means (1) that because Melissus accepts any atopos, all other atopoi follow; or (2)
that it’s because Melissus accepts a very specific atopos, namely, that everything is one, all other
atopoi follow. The former is far more suggestive of explosivity than the latter.
27
ἔsi ἤsoi pEqὶ ἅpamsa oὕsx1 ἔvEi, jaὶ ἔrsi jaὶ kEtjὸm jaὶ oὐ kEtjὸm jaὶ ὂm jaὶ oὐj ὄm, jaὶ
pEqὶ sὰ1 ἄkka1 u arEi1 jaὶ ἀpou arEi1 ὁloiosqόpx1 [5df ALL], ἢ oὒ ἀkkὰ pEqὶ lem sima1 [5df
SOME], pEqί sima1 d᾽ oὔ. jaὶ Eἰ lὲm lὴ pEqὶ p ara1, aὗsai ἂm EἶEm ὁlokogoύlEmai: Eἰ dὲ pEqὶ
para1. . . (1008a8–11). Up until this point I have been relying exclusively on the English transla-
tion of the Metaphysics in the Barnes. J. Barnes, The Complete Works of Aristotle, Volumes I and II
(Princeton, NJ: Princeton UP, 1984). But that translation of this crucial passage seems to insert
some English words that do not correspond to the Greek in any obvious way, e.g., “compatible” and
“true”. Kirwan’s translation of 1008a8–11 does not use “compatible” or “true”: “Again, either this
is so in every case, i.e. a thing is both pale and not pale, both a thing-that-is and not a thing-that-is,
and in a similar way for all other assertions and denials; or it is so in some cases but not in others. If
it is not so in all cases, these would be agreed. But if it is so in all [. . .]” (Kirwan 1993: 13).
28
Kirwan (1993): 104.

45
 DANIEL COREN

about the matter.29 My view is that Aristotle’s dismissal of SOME is

“inadequate” if and only if a more thorough treatment of the position is required.
But it’s not clear that such a treatment is required. In Section VII, I hope to show
why this is so.
I take this passage as evidence that Aristotle does not think that SOME entails (or
explodes into) ALL. On the supposition that he did think that SOME explodes into
ALL, it’s very difficult to see why Aristotle would have stated that, in the case of
SOME, the exceptions—those contradictions that are not true—would be agreed
upon (homologe^ o). He would have stated that the exceptions could not be agreed
upon, since, on this supposition, claiming that there are some violations of PNC
entails that all contradictions are true. It would not be inconsistent, though, for Aris-
totle to argue that SOME does not explode into ALL while also arguing that a falsifi-
cation of ALL provides a very good reason for thinking that SOME is false.
Finally, Aristotle’s neat distinction between SOME and ALL at 1008a8–11 does
not, in at least one important respect, come as a surprise. Aristotle was the first
thinker to carefully articulate the distinction between a pair such as SOME and
ALL, namely, a contradictory (antiphatikos) as opposed to a contrary (enantios):
I call an affirmation and a negation contradictory [ἀmsiuasijό1] opposites when what one signifies
universally the other signifies not universally, e.g. every man is white – not every man is white, no
man is white – some man is white. But I call the universal affirmation and the universal negation
contrary [ἐmamsίo1] opposites, e.g. every man is just – no man is just. So these cannot be true
together, but their opposites may both be true with respect to the same thing, e.g. not every man is
white – some man is white (De Int.17b.17–26).

We find the same distinction expressed in a slightly different way in Aristotle’s

Metaphysics I (X): “Among opposites, contradictories admit of no middle term;
for contradiction is this—an opposition, one or the other side of which must
attach to anything whatever, i.e. which has no intermediate” (1057a33–35). Aris-
totle was acutely aware of the fact that contradictories exhaust the logical space,
admitting of no intermediates, whereas contraries do not exhaust the logical
space (and therefore admit of intermediates). It seems highly unlikely, then, that
Aristotle would have failed to recognize the fact that “All contradictions are true”
is the contrary of “No contradictions are true,” or the fact that “Not all contradic-
tions are true” is the contradictory of “All contradictions are true.”30 So it also
seems highly unlikely that Aristotle would have failed to recognize that proving
29
Politis (2004: 139, 153) also notes the distinction, and has more to say about it than Kirwan. This
will be taken up in Section VII. The distinction is also noted and briefly discussed in Gottlieb
(1994).
30
However, as Bailey (D. T. J. Bailey, “Plato and Aristotle on the Unhypothetical,” Oxford Studies in
Ancient Philosophy 30 (2006): 101–26) points out, Aristotle formulates PNC in terms of contraries
rather than contradictories in what I have as [B].

46
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

that the contrary of PNC is false does not entail that the contradictory of PNC is
false.31

V. ONE VITAL REASON ARISTOTLE GIVES SO MUCH ATTENTION TO

ALL: HE TOOK HIS RADICAL, REAL OPPONENTS TO HOLD VIEWS
DEEPLY CONNECTED TO ALL

Throughout his dense argumentation in defense of PNC in his Metaphysics

C.4, there are only two instances where Aristotle mentions specific opponents.
He writes that Anaxagoras holds the view that “all things are mixed together; so
that nothing really exists,” and Protagoras the view that “it is equally possible to
affirm and to deny anything of anything” (C.4 1007b21–22, 25–26). Other than
that, the arguments in C.4 seem to be aimed at imaginary opponents. But in C.5,
Aristotle’s opponents become much more concrete. He discusses views he takes
to be held by Anaxagoras, Cratylus, Democritus, Empedocles, Heraclitus,
Parmenides, Protagoras, and Xenophanes. Even Homer is credited with holding
an anti-PNC view.
Aristotle begins by arguing that since Protagoras holds that “all opinions and
appearances are true” (a view discussed in Plato’s Theaetetus as the doctrine,
“man is the measure of all things”), it follows that Protagoras is also bound to
accept the view, “all statements must be at the same time true and false”
(C.5 1009a7–9). For it seems that people disagree about a great many matters,
and each person is convinced that his or her opinion is correct. If Protagoras is
right, then each pair of contradictory opinions is true. Each contradiction is true.
The subject of each opinion both has the predicate asserted in that opinion and
also has the predicate asserted in the corresponding contradictory opinion
(1009a10–16). So Aristotle thinks that Protagoras’s view entails the contrary of
PNC (ALL).32 Perhaps Aristotle is a little too quick here: for one thing, it seems
rather generous to presuppose that there does not exist a single thing about which
people do not hold both an opinion and its contradictory. For (1) there are pre-
sumably things we have not had a chance to form opinions about yet, and (2)
31
There is also a good deal of evidence that Aristotle wrote an entire treatise on contrariety, though it
did not make its way down to us. See Diogenes Laertius V, 21 (30); Hesychius in onomatologos (ch.
De. Ar. Libris, p.49), s.v. (30) pEqὶ ἐmamsίxm ᾶ; Simplicius, in categ. f.t. 3a (p. 83 b7); chiefly see
Aristotelis Fragmenta, collected by Valentinus Rose, pp. 109–14, 118–24. V. Rose (ed.), Aristotelis
Fragmenta (Leipzig, 1886). This evidence makes it even more unlikely that Aristotle would have
confused contradictories with contraries, let alone in such a central and drawn-out discussion.
32
As Lee (2005: 120) explains, this argument of Aristotle’s “is in fact based on Plato’s self-refutation
arguments in the Theaetetus, where Plato argues that Protagoras is committed to thinking that some
beliefs are true and false and thereby refutes Protagoras’ original claim that all beliefs are true”.
Aristotle draws the further implication, “everything both is and is not F, and vice versa”.

47
 DANIEL COREN

there might be some subjects of opinion on which people do not disagree. But it
still seems reasonable for Aristotle to infer that Protagoras’s view and its impli-
cations fall much closer to ALL than the comparatively moderate SOME.
Aristotle goes on to argue in C.5 that thinkers such as Anaxagoras, Cratylus,
Democritus, Empedocles, Heraclitus, and Xenophanes have been led into confu-
sion about matters as fundamental as the truth of PNC because of “observation of
the sensible world” together with the assumption that knowledge is sensation,
without appropriately qualifying their descriptions and understanding of what
goes on in the sensible world (1009a22–23). Some of them, such as Anaxagoras,
observe “contraries coming into existence out of the same thing”; relying on the
assumption that something that already exists cannot again come into existence,
Anaxagoras arrives at the view that things “must have existed before as both con-
traries alike”, with everything “mixed in” with everything else (1008a23–27). A
little later on, Aristotle attributes to Anaxagoras a doctrine that seems to be the
same as Protagoras’s: “things would be for [them] as they supposed them to be”
(1009b27–28). Aristotle explains that others, such as Democritus, hold that “the
void and the full exist alike in every part, and yet one of these is being, and the
other non-being” (1009a28–30). So being and non-being exist in every part of
everything. These sound like radical views that Aristotle takes to be close to—if
not identical with—ALL.
Aristotle seems to think that Empedocles holds a similar view to those held by
Democritus and Anaxagoras. As evidence, he claims that “Empedocles says that
when men change their condition they change their knowledge”; that “Wisdom
increases in men according to their present state”; and “So far as their nature
changes, so far to them always come changed thoughts into mind” (1009b17–
22). It’s unclear whether—and if so, how—Aristotle thinks that any view
expressed in these quotations is the same as, for example, Anaxagoras’s
“everything is mixed together with everything” view, let alone Protagoras’s doc-
trine that each thing in the world is exactly as it appears to each person, even if
there are many instances of contradictory opinions (about the same object, in the
same respect, at the same time). For Empedocles might maintain that I am wise
in the morning and not wise in the afternoon, without being in the slightest dan-
ger of conceding that I am both wise and not wise at one and the same time.
More charitably, we probably ought to take careful note of Aristotle’s lumping
Empedocles in with those “thinkers [who] suppose knowledge to be sensation,
and this to be a physical alteration [. . .] [so that] what appears to our senses must
be true” (1009b13–15). So what Aristotle seems to mean here is that Empedocles
allows that one and the same thing might be sensed by two (or more) people
simultaneously; the thing might be sensed as cold by one person and as hot by
another. So, if knowledge just is sensation, as Anaxagoras, Democritus, and
Empedocles all hold (according to Aristotle) then one and the same the thing is

48
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

at the same time both known to be hot and known to be cold. Aristotle seems to
have in mind the further inference that if knowledge is something like justified
belief that is “tied down” to the truth, as is probably the most promising defini-
tion of knowledge discussed in Plato’s Meno, then it would be true that a single
object is at the same and in the same respect both hot and cold. That’s a contra-
diction. Aristotle seems to think that this generalizes: there’s an endless prolifera-
tion of contradictions. So we get ALL.
Parmenides apparently speaks “in the same way” (sὸm aὐsὸm sqόpom)
(1009b22–25):
For as in each case the much-bent limbs are composed
So is the mind of men; for in each and all men

‘Tis one thing thinks – the substance of their limbs;

For that of which there is more is thought.

Aristotle takes Parmenides in this passage to be asserting that thought or intellect

in each human being is merely a product of the way the human body is com-
posed, not differing at all from the way in which human senses are composed.
Why should this entrench Parmenides’s views with those of Anaxagoras,
Democritus, Empedocles, and Protagoras? Aristotle seems to have something
like the following reasoning in mind: if knowledge just is sensation, and if what
is sensed is relative to the way in which that person is composed, then knowledge
will be relative to the way in which that person is composed. But if knowledge
requires truth, and since people often have contradictory opinions or impressions
of one and the same thing, then, Aristotle seems to worry, it will be both true and
false to say that one and the same object is in some state. Jane might “know” that
some pile of sand is hot and, at the very same time, John might “know” that the
very same pile of sand is cold. Of course, from this quotation alone it’s not clear
why Aristotle does not consider that Parmenides could maintain that things in the
world might not admit of any contradictions, but simply that human beings can-
not guarantee epistemic access to the way things really are. So Parmenides might
hold, as far as Aristotle’s provided evidence goes, that Protagorean relativism is
true as far as human beings are concerned but perhaps false in reality. Aristotle
seems to think, though, that Parmenides is committed to the stronger claim that
Protagorean relativism is true not merely for us but also in reality; everything is
or is not according to the way each of us is composed. So, Aristotle appears to
infer, Parmenides thinks that everything both is and is not. Thus, again, we get
ALL.
Cratylus, Heraclitus, Homer, and Xenophanes are also credited in Aristotle’s
Metaphysics C.5 with radical positions that would make it “natural that beginners

49
 DANIEL COREN

in philosophy should lose heart” (1009b36–37). The evidence Aristotle gives for
Homer’s anti-PNC view is decidedly obscure (1009b29–32):
He made Hector, when he was unconscious from the blow, lie ‘thinking other thoughts’, which
implies that even those who are bereft of thoughts have thoughts, though not the same. Evidently,
then, if both are forms of thought, the real things also are at the same time so and not so.

Hector himself is not obscure. He’s a key figure in Greek mythology and, more
pertinently, in, for example, Homer’s Iliad. But Aristotle leaves it unfortunately
abstruse just what exactly is the connection between this description of Hector’s
“thinking other thoughts” and the view that all contradictions are true or the view
that everything is as it appears to each of us. Why should not Homer be writing
figuratively when he says that Hector “thinks other thoughts” when rendered
unconscious? Even if we grant that Homer must be taken literally, why is
Aristotle so quick to assume that unconscious people cannot have dreams, and
that dreams do not count as “thinking other thoughts”? If indeed all of these
assumptions are granted, it’s still not clear why Aristotle thinks that we get any
contradictions outside “Hector while unconscious was both thinking and not
thinking in the same time, in the same respect, and with all further qualifications
added to meet dialectical objections.” But probably Aristotle wants the reader to
treat this as one sample among a vast array of either explicitly contradiction
endorsing or Protagorean relativism endorsing—and, implicitly, contradiction
endorsing—passages and stories.
Aristotle also considers Heraclitus’s view that the world is changing every-
where and in every respect, such that “no true statement can be made”
(1010a6–9). Part of Aristotle’s diagnosis of what has gone wrong is by now
familiar: Heraclitus thinks that nature is all that there is, and what can be per-
ceived in nature is all that can be known. But Heraclitus adds the doctrine that
there is only movement and change in nature, in every respect. It’s not clear pre-
cisely if and how Aristotle thinks that this Heraclitean view entails that all con-
tradictions are true. For if no true statement can be made then presumably
Heraclitus would not endorse the truth of the statement, “All contradictions are
true.” Superficially, Heracliteanism and ALL are at odds with one another: if ALL
is true then every statement is true, whereas if Heracliteanism is true then, appar-
ently, no statement is true. Though he does not make this explicit, perhaps
Aristotle sees a deeper connection between ALL and Heracliteanism. Both posi-
tions entail that “to seek the truth would be to pursue flying game”
(1009b37–38). They make truth impossible to pin down or understand.
Heracliteanism does so by insisting that every candidate in the world for a true
statement is constantly changing in every way, and ALL does so by insisting that
everything that could be said about anything in the world in any way is true (and
false). Aristotle seems to think that a proponent of such a radical view will end

50
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

up like poor Cratylus, “who finally did not think it right to say anything but only
moved his finger, and criticized Heraclitus for saying that it is impossible to step
twice into the same river; for he thought one could not do it even once”
(1010a12–15).
Aristotle’s opponents are, then, a far cry from those contemporary philoso-
phers who reject PNC only for special cases such as certain semantic paradoxes.
His opponents do not selectively reject PNC. They annihilate PNC. By Aristo-
tle’s lights, at least, opponents such as Anaxagoras, Cratylus, Democritus, Hera-
clitus, Parmenides, and Protagoras do not just posit some select areas where there
are or might well be exceptions to PNC. Rather, they take being and non-being to
be everywhere and in everything simultaneously; they argue that everything is
changing constantly and in every respect; they hold that everything is mixed with
everything else; they assert that all opinions, including all contradictory opinions,
are simultaneously true. In this historical context, we can make much more sense
of Aristotle’s apparently exclusive focus on PNC’s contrary rather than PNC’s
contradictory. From a strictly logical perspective, PNC’s contrary, ALL, might
well be thought largely irrelevant if Aristotle aims to defend PNC. But from a
more concrete, historical perspective, Aristotle’s failure to directly address
SOME makes a good deal of sense.

VI. A SECOND IMPORTANT REASON ARISTOTLE’S DEFENSE OF

PNC SEEMS EVASIVE: HE IS CONFINED TO AN INDIRECT
DEFENSE OF PNC

For Aristotle, the truth of PNC is better known than anything else, and is more
certain than anything else. Yet PNC cannot be directly shown. It cannot be shown
in a proof or a demonstration. There are at least three reasons why. First, Aris-
totle means something very specific by a “demonstration” (apodeixis). In a dem-
onstration, the premises must be more certain than the conclusion. Since nothing
can be more certain than the PNC, it follows necessarily that PNC cannot be the
conclusion of a demonstration. Second, the truth of PNC, as the most certain
principle of all principles, is at least implicitly relied upon in the premises of
every demonstration. It follows, then, that any demonstration with PNC as a con-
clusion would have premises in which the truth of PNC is at least implicitly
relied upon. So the demonstration in question would employ viciously circular
reasoning or, more colloquially, it would commit the fallacy of begging the ques-
tion against his opponents.
A third reason runs as follows: (1) It’s impossible that there could be a demon-
stration of “absolutely everything,” since otherwise there “would be an infinite
regress” (Metaph.C.4.1006a8). In particular, if it were the case that absolutely

51
 DANIEL COREN

everything must be demonstrated, including all premises used to prove less cer-
tain (bebaios) conclusions in all demonstrations, and especially those demonstra-
tions used to prove archai, then there would be a series of demonstrations that
has no end and thus is infinite (apeiron).33 In that series, the things being used
currently as premises to demonstrate those things used previously as premises—
to prove things like first principles—must themselves be the conclusions of fur-
ther demonstrations, and so on ad infinitum. That would be absurd, on Aristotle’s
account, because “there would still be no demonstration” (Meta-
ph.C.4.1006a9)34; (2) Given (1), it follows that the series of demonstrable things
is finite. Since the series is finite, it has a beginning. At that beginning, there are
indemonstrable things. Those indemonstrable things are first principles: they are
starting points for all demonstrations; and (3) given that the PNC is more certain
than anything else, including all other first principles, and given (1) and (2), it
follows that the PNC must be one of those indemonstrable things
(Metaph.C.4.1006a5–11). In fact, the PNC is more certain than any other first
principle to be one of those indemonstrable things. Nothing is presupposed more
often in demonstrations than the PNC, and nothing is more likely to be
indemonstrable.35

33
As is noted in Lear (J. Lear, Aristotle: The Desire to Understand (Cambridge: Cambridge UP,
1988): 251), “There are two conditions which a principle must satisfy if it is to be the most certain
of all. First, it must not be possible to be mistaken about it. Second, anyone who understands any-
thing understands the principle.” The first condition is from 1005b12, the second from 1005b16.
34
Aristotle does not spell out exactly why he thinks that an infinite series of demonstrations entails
that there would be no demonstration. But he might have in mind something like the following: Sup-
pose one proves that “All A’s are C’s” using two premises: “All A’s are B’s” and “All B’s are C’s.”
“All A’s are B’s” must be proved, as must “All B’s are C’s.” Suppose both premises are proved.
Then the premises used to prove both premises must themselves each be proved. And the process
iterates an infinite number of times. Further, for Aristotle “something is infinite if, taking it quantity
by quantity, we can always take something outside” (Physics III.6.207a7–8). So call each stage of
the iteration a quantity. For any quantityn, there is a quantity outside quantityn, namely, quantityn11.
The premises of quantityn are proved in quantityn11, but the premises in quantityn11 are left
unproved. Also, see Aristotle’s Posterior Analytics I 3.72b5–24, where he argues that it’s correct to
think that if “we are led back forever” in our demonstrations of what is prior then “we would not
understand what is posterior. . .for it is impossible to go through infinitely many things.”
35
Whitaker (C. W. A. Whitaker, Aristotle’s De Interpretatione (Oxford: Oxford UP, 1996)) notes that
there is an important difference between a positive demonstration and a negative demonstration,
and that C.4 employs only the latter. A negative demonstration “does not give reasons for the truth
of the principle, or an understanding of it, either of which would be impossible; rather it gives rea-
sons why the opponents cannot hold their position, and an understanding of why they must be
wrong. If the opponents are wrong, then the principle is true” (Whitaker 1996: 188). Wedin (M. V.
Wedin, “Aristotle on the Firmness of the Principle of Non-Contradiction,” Phronesis 49 (2004a):
225–65) contains an excellent discussion of Aristotle’s strategy in showing that PNC is not just any
old indemonstrable principle but instead the firmest of all indemonstrable principles.

52
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

This forces Aristotle’s defense of PNC into a rather unusual and delicate
position. As Terence Irwin helpfully explains, Aristotle’s defense must be
“neither demonstrative nor purely dialectical.” For if Aristotle’s defense is
demonstrative, then we run into familiar problems such as using PNC to dem-
onstrate PNC. If, on the other hand, the defense is purely dialectical then
there is no guarantee that the conclusion will be “knowledge of the truth” as
opposed to “a coherent account of common beliefs.”36 Moreover, using what
we today would call a reductio ad absurdum argument, an argument to the
effect that the opponent’s position entails inconsistency, is much too clumsy.
For a reductio will not trouble an opponent who thinks that contradictions are
true: “this is a peculiar case in which forcing your adversary to concede the
contradictory of his own thesis is not a clear-cut dialectical trump.”37 Aristotle
would seem to obviously beg the question against his anti-PNC opponents if
he claimed that their position is untenable because their position entails a
rejection of PNC. So if Aristotle tries to use the Socratic “elenchus” method
of argument, wherein Socrates guides his opponent toward the admission that
his opponent’s position entails an inconsistency (described in so many of the
dialogues that populate Plato’s extant writings), this technique must be han-
dled with special care. It cannot be applied in a straightforward way and one
might worry, it can do no better than helping to persuade a neutral arbiter—as
opposed to Aristotle’s contradiction-welcoming opponents—of the truth of
PNC. As I’ll argue in VII, though, Aristotle aims at convincing not just a
neutral arbiter but also his contradiction-welcoming opponents that coherent
communication in general, including his opponent’s communication of his or
her anti-PNC position relies on non-contradiction.

VII. WHY ARISTOTLE’S DEFENSE OF PNC UNDERMINES NOT JUST

ALL BUT ALSO SOME

Why does Aristotle take his defense of PNC to be successful, despite his
appearing to be exclusively interesting in attacking ALL (“All contradictions are
true”) rather than SOME (“Some contradictions are true”)? Or, distracted by a
plethora of predecessors with confused and radical views, and confined to a par-
ticular kind of indirect means of defense, has the task of defending a principle
more certain and foundational than any other “reduced Aristotle to sheer

36
T. H. Irwin, Aristotle’s First Principles (Oxford: Oxford UP, 1988) 180, especially ch. 9, sections
98–106.
37
Castagnoli (2010): 74.

53
 DANIEL COREN

babble”?38 I think that Aristotle was not reduced to sheer babble. Instead, it
seems to me that Aristotle’s defense of PNC can be read in a way that is not just
excusable and sensible in its historical and dialectical context, but also in a way
that is more objectively, philosophically attractive. Against the PNC-rejecter,
especially the one who demands that everything, including PNC, must be demon-
strated directly, Aristotle begins his defense of PNC with an exclusive disjunc-
tion: Either the opponent says something (the indefinite si) or nothing (lhuem,
m^ethen) (1006a12–13). If the opponent says not something (si) but, instead,
nothing, then Aristotle’s opponent cannot reason about anything at all. That is
because, on Aristotle’s account, saying something—and as we will see momen-
tarily, something that signifies in a particular way—is a necessary condition for
reasoning or communicating about anything at all. Not only is such a person
incapable of reasoning or communicating with others, but also such a person is
incapable of being reasoned or communicated with by anyone else, including
herself. Similarly, plants do not reason at all. Nor can anyone or anything else
reason with plants. Even plants cannot reason with plants. So Aristotle concludes
that if his opponent says nothing then his opponent is “no better than a mere
plant” (1006a15).
But what if Aristotle’s opponent decides to be a little more obliging, so that
she says something? It turns out that Aristotle does not require his opponent to be
very obliging at all, for by “something” Aristotle does not mean that the oppo-
nent must “say that something either is or is not [Eἶmaί si kegEim ἢ lὴ Eἶmai]”
(1006a16–17). Such a request, Aristotle notes, might beg the question against his
opponent. For saying that something either is or is not would presuppose the truth
of PNC. But the truth of PNC is precisely the point that needs to be proved by
Aristotle against the PNC-rejecter. Rather, by “something” Aristotle means
“something which is significant [so that it signifies something] both for himself
[the opponent] and for another [ἀkkὰ rhlaίmEim ge si jaὶ aὑsῷ jaὶ ἄkkῳ]”
(1006a21–22). “To signify” (s^ emainein) something by what one says is taken by
Aristotle as a necessary condition for saying “anything” (1006a23). That’s
because to signify nothing by what one says is to be in a situation similar to the
38
Anscombe (G. E. M. Anscombe, “Modern Moral Philosophy,” Philosophy 33 (1958): 1–19)
famously argued that “[the difficulty of the concept ‘pleasure’] reduced Aristotle to sheer babble
about ‘the bloom on the cheek of youth’ because, for good reasons, he wanted to make it out both
identical with and different from the pleasurable activity.” There are two notable connections
between my discussion here and Anscombe’s: First, making pleasure both identical with and not
identical with the pleasurable activity would be a violation of PNC (so long as the context in which
it is identical with the pleasurable activity and the context in which it is non-identical are in all
respects the same). Second, if Aristotle’s defense of PNC fails to defend a claim anywhere near as
strong as PNC, without any further justification for that move, then it’s difficult to see how Aristotle
is not babbling.

54
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

plant-like opponent who chose to say nothing in the first place. Aristotle requires
his opponent to say something that’s not just significant for the opponent but also
for another. If the PNC-rejecter is required to say something that is significant for
the PNC-rejecter but not necessarily also for another, then she might say some-
thing that fails to signify anything for another. But such a thing might be nonsen-
sical. For example, to say “woman” is likely to be significant for myself and for
another, but to say “shabadabadoo” is not.39
Once his opponent has said something that is significant both for the opponent
and for another, “we shall already have something definite [h^ orismenon]”
(1006a24–25). Aristotle argues that if I say something by which I signify some-
thing for myself and for another—which we will call the s^ emainein condition—
40
then I have conceded that at least one thing is determinate. The result of so con-
ceding, Aristotle argues, is that “something is true apart from demonstration”:
“not everything will be ‘so and not so’ [oὐj ἂm pᾶm oὕsx1 jaὶ oὐv oὕsx1 ἔvoi]”
(1006a27–28). Aristotle’s first argument in defense of PNC (at 1006a12–28) can
be stated in slightly more general terms as follows.
Premise 1: A person (such as a PNC-rejecter) may either say something, or nothing.
P2: If a person (such as a PNC-rejecter) says nothing, she cannot reason with anyone. Nor can
anyone reason with her.

P3: To say “something” is to say something that is significant to oneself and to another – the
s^emainein condition.
P4: If something satisfying the s^emainein condition is said, then at least one thing is determinate.
P5: If at least one thing is determinate then not everything is “so and not so”. (In particular, the one
thing from P4 is not “so and not so”).

39
It is not entirely clear to me what exactly Aristotle means by jaὶ ἄkkῳ (“and another”) in
rhlaίmEim ge si jaὶ aὑsῷ jaὶ ἄkkῳ at 1006a21–22. Who is the other? Does Aristotle mean
“another [person]” in the sense of a person arbitrarily selected from all people, or does he mean a
person non-arbitrarily selected from those people familiar to him or to his opponent? My suspicion
is that Aristotle has something closer to the arbitrary selection in mind for the jaὶ ἄkkῳ condition,
mainly because the request as a whole is not to say something that Aristotle has in mind, let alone
something Aristotle has suggested. Aristotle is giving his opponent his pick of just about the broad-
est range of options: any particular sound that signifies something “for you and someone else.”
40
Just as Aristotle seems to equate radical indeterminacy with everything being “so and not so,”
Aristotle also seems to equate the possibility that one word has an infinite number of meanings with
the possibility that one word has no meaning at all: “If, however, [the meanings assigned to a word]
were not limited but one were to say that the word has an infinite number of meanings, obviously
reasoning would be impossible; for not to have one meaning is to have no meaning [. . .] [Eἰ dὲ lή
sEuEίh, ἀkk᾽ ἄpEiqa rhlaίmEim uaίh, uamEqὸm ὅsi oὐj ἂm Eἴh kόgo1: sὸ gὰq lὴ ἓm rhlaίmEim
oὐuὲm rhlaίmEim ἐrsίm]” (1006b6–9).

55
 DANIEL COREN

Conclusion: Either the PNC-rejecter cannot be reasoned with and no one can reason with him, or he
concedes that not everything is “so and not so”.

In Aristotle’s argument above, as I have reconstructed it, I take Premise 4 to be a

very strong claim. It seems to require that to say anything that satisfies the
s^emainein condition must concede that there is something that is determinate.
Suppose it’s not the case that Aristotle’s argument requires P4 to be as strong as
“Anything that can be said by anyone (that also satisfies the s^emainein condition)
concedes that something is determinate.” So a weaker claim will work. For
example, change P4 to “Almost anything that can be said by anyone (that also
satisfies the s^emainein condition) concedes something that is determinate.” The
truth of such a claim is consistent with there being at least one thing satisfying
the s^emainein condition that can be said by anyone, but which does not concede
something that is determinate. It is possible, then, that when Aristotle asks his
opponent to say something satisfying the s^emainein condition, his opponent says
that very thing—the thing that satisfies the s^emainein condition but without con-
ceding something that is determinate. But then P4 is false.
I think that the premises in the argument above might well prove a conclusion
a good deal stronger than just “Not everything is ‘so and not so’.” They seem to
show that there are a great many things that are determinate. In fact, the premises
go a long way toward disproving SOME and defending PNC. There are, after all,
a great many ordinary things that are said in the world that would seem to safely
satisfy the s^emainein condition: a great many things signify something (as
opposed to nothing, and as opposed to everything) for the person saying it and
for another. It’s in virtue of the fact that different words and different combina-
tions of those words tend to each mean different, defined and determinate
things—both to the person saying it and to other people—that Aristotle’s oppo-
nent understands (1) what Aristotle means when he makes the sounds forming
the words expressing PNC and (2) why those sounds made by Aristotle express
something, namely, PNC, that contradicts the opponent’s own view. For if Aris-
totle’s opponent did not understand the sounds made by Aristotle as expressing
PNC (and not the contradictory of PNC), then it’s unlikely that there would be
any opponent with whom Aristotle has to argue. More generally, take every word
that has been used in every conversation involving at least two human beings in
the history of the world. That includes, presumably, a very large amount of
words. It seems a broad but fairly uncontroversial assumption that at least a sig-
nificant percentage—perhaps even the vast majority—of those words satisfy the
s^emainein condition. After all, it’s unlikely that conversations would have con-
tinued, and languages developed over the centuries if it were not the case that in
the vast majority of those conversations, what was said was understood more or
less as intended.

56
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

Further, even many sounds that are not words tend to each have different
significations satisfying the s^emainein condition. A groan of displeasure is not
a word. But a groan of displeasure seems to be understood as meaning more
or less the same thing each time it is heard and made, even among people
who speak different languages. So a groan of displeasure seems to satisfy the
s^emainein condition.41 So, too, an exclamation as a result of a shock or sur-
prise is a sound that very often is not a word at all. Yet, if I were to make
such a sound right now, I suspect that the people around me would think that
I had just been shocked or surprised—perhaps they’d also be a little shocked
themselves. But in particular, they probably would not take my sound to indi-
cate that I am bored or, more generally, not shocked. It’s a similar case, I
think, with laughter: laughter is just a sound rather than a word, but seems to
satisfy the s^emainein condition. Indeed, it appears that many other types of
sounds are very likely to satisfy the s^emainein condition.42 These seem to be
the sorts of common sense assumptions, inferences and intuitions Aristotle has
in mind.
It appears, then, that there are a great many things which are not “so and not
so,” according to the premises of the argument at 1006a12–28. Here is why: (1)
there a great many things that satisfy the s^emainein condition; (2) saying any-
thing satisfying the s^emainein condition concedes something that is determinate;
(3) anything that is determinate falsifies “so and not so”; (4) anything that is not
“so and not so” is a falsification of ALL; (5) there are evidently a great many
things that are not “so and not so”; and therefore (6) there are a great many things
that falsify ALL. In fact, there seem to be so many such falsifications that I’m
inclined to think that, taken together, all those falsifications come at least very
close to falsifying SOME. Thus the premises in the argument at 1006a12–28, the

41
Perhaps a case can be made for the claim that satisfaction of the s^emainein condition extends even
beyond sounds to signs. After all, a frown seems to have a meaning more or less universally distinct
from that of a smile. Similarly, raising one’s eyebrows might satisfy the s^emainein condition. But
here I may be straying from Aristotle’s argument at 1006a12–28, since Aristotle states that the
opponent’s silence—saying lhuem as opposed to si—rules out the opponent’s ability to reason or
be reasoned with.
42
Aristotle notes elsewhere, “Even inarticulate noises (of beasts, for instance) do indeed reveal some-
thing” (De Int.16a 28–29). J. Barnes, The Complete Works of Aristotle, Volume I (Princeton, NJ:
Princeton UP, 1984). Perhaps it is possible on Aristotle’s account that even sounds made by non-
human animals can satisfy the s^emainein condition. After all, animals mean something by the
sounds they make, and other animals often understand that meaning. Even the ancient sceptics
seemed to be thoroughly convinced of this point: “We certainly see animals uttering quite human
cries – jays, for instance, and others” (PH i 74). R. G. Bury, trans., Sextus Empiricus (Loeb Classical
Library), 4 volumes (Cambridge, MA: Harvard UP, 1939–49).

57
 DANIEL COREN

focal point of which is Aristotle’s theory of what counts as a thing that is deter-
minate, seem to come at least very close to defending PNC.43
In fact, one might well think that Aristotle’s defense of PNC is more effective
against a moderate (if also imaginary) opponent, one who holds that some but not
all contradictions are true, than against a much more radical opponent such as
Anaxagoras or Protagoras. For an opponent who holds that it’s the case that (8x)
(8F)(Fx & Fx) can accept every premise and inference in Aristotle’s arguments
without having her position undermined. Each of those premises and inferences
might simply be a conjunct in one of the conjunctions forming a small sample of
the contradictions that Aristotle’s opponent takes to be true. His opponent also
rejects every premise and inference in Aristotle’s arguments: those are just the cor-
responding conjuncts in that small sample. She can maintain that every one of Aris-
totle’s claims is question-begging (and that every one of his claims is also not
question-begging). Of course, Aristotle might point out that if this is so then it’s not
at all clear why his opponent has bothered to articulate an anti-PNC position or, for
that matter, any position at all. For if his opponent’s position is indeed that radical
then by his opponent’s lights it’s also the case that PNC is true, indemonstrable, and
more certain than any other first principle. But the dialectic between Aristotle and a
proponent of (9x)(9F)(Fx & Fx) is more manageable. For if Aristotle shows—as
I have argued he can and does—that the necessary conditions for coherent commu-
nication militate against the conditions for contradictions to be true, then a propo-
nent of SOME, (9x)(9F)(Fx & Fx), has a good reason to severely weaken her
position (from “there are some true contradictions” to “it’s possible that there is
some small pocket of true contradictions as yet undiscovered”) if not to give up her
position entirely, and to whole-heartedly reject ALL.

VIII. CONCLUSION

I mentioned in Section III that Lukasiewicz44 seems to be the first to have

noted that Aristotle’s defense of PNC appears to defend PNC against its
43
Gottlieb (1994) argues that Aristotle’s defense of PNC employs a strategy that is the reverse of Pla-
to’s strategy in attacking Protagorean relativism in the Theaetetus. That is, on her account, Aristotle
is arguing that because it is the case that meaningful communication, including meaningful lan-
guage as we know it, is indeed possible, it follows that PNC-rejecters such as Protagoras must hold
not just that “Some things are so and not so” but that “Everything is so and not so.” As a result, “if
Aristotle and Plato are right, a Protagorean would have to reject PNC across the board, so Aristotle
is not addressing a straw opponent nor is he confused when he argues against someone who holds
the contrary and not the contradictory of PNC” (Ibid). But as I show here, I think that we can read
Aristotle as showing not merely that his opponents’ views are deeply connected with—if not simply
entailing—ALL, but, more directly, as using the foundations for coherent communication to seri-
ously undermine SOME (rather than just ALL).
44
Lukasiewicz (1910).

58
 WHY DOES ARISTOTLE DEFEND THE PRINCIPLE OF NON-CONTRADICTION

contrary—ALL—rather than its contradictory—SOME. The closing sentences of

Lukasiewicz’s paper are also worth quoting:
Denial of the principle of contradiction would have opened door and gate to every falsity and
nipped the young, blossoming science in the bud. Hence, the Stagirite turns against the opponents
of the principle with forceful language in which one can trace an internal fervor, against the eristic
thinkers of Megara, the cynics of the school of Antisthenes, the disciples of Heraclitus, the partisans
of Protagoras; and he battles with all of them for a theoretical principle as if for personal goods. He
might well have himself felt the weakness of his argument, and so he announced his principle as a
final axiom, an unassailable dogma.

I have tried to show that the aforementioned Stagirite, no doubt with some inter-
nal fervor, gives a defense of PNC that is (1) aimed at real, radical opponents
who Aristotle took to hold views deeply connected to PNC; (2) must be a particu-
lar kind of indirect defense than a direct demonstration; and (3) demonstrates the
reliance of coherent communication on non-contradiction. Read this way, Aristo-
tle’s defense of PNC is not just historically and dialectically excusable and sensi-
ble, but also philosophically attractive. He gives a fairly compelling case to
reject not just PNC’s contrary but also PNC’s contradictory.45

University of Colorado

45
For helpful comments on earlier drafts of this paper, I am grateful to Dominic Bailey, Paula
Gottlieb, Kathrin Koslicki, Mitzi (Mi-Kyoung) Lee, Robert Pasnau, and an anonymous reviewer for
the Philosophical Forum.

59
Copyright of Philosophical Forum is the property of Wiley-Blackwell and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.