Professional Documents
Culture Documents
Da Costa & Wolf - Studies in Paraconsistent Logic I. The Dialectical Principle of The Unity of Opposites
Da Costa & Wolf - Studies in Paraconsistent Logic I. The Dialectical Principle of The Unity of Opposites
189
NEWTONC.A. DA COSTA and ROBERTG. WOLF
190
STUDIES IN PARACONSISTENTLOGIC I
191
NEWTONC.A. DA COSTA and ROBERTG. WOLF
any way contrary to classical logic and hence does not call for any
revision in logic. They also indicate that the terms "forces", "move-
ments" and "tendencies" are systematically ambiguous. What counts
- say - as a "force" varies quite radically from science to science. A
"force" at work in a gravitational system is quite different from a
"force" at work in a political system. The dialectical principle is
meant to cover all cases.
This raises the immediate question as to the status of the prin-
ciple of the unity of opposites. McGill and Parry interpret it as "a
philosophical generalization of findings in various sciences." (McGill
and Parry 1948, p. 426) They go on to suggest that its usefulness
might lie in giving "useful guidance to a scientist engaged in scien-
tific research, especially in suggesting possible correctives to one-
sided emphases in current scientific theories." They conclude that
"the principle of the unity of opposites may therefore be said to ex-
press methodological experience, and revised judgments, in several
fields of science." (McGill and Parry 1948, p. 427)
Without doubting that such a principle would arise as (i.e., be
psychologically suggested by) a generalization of practice in various
sciences and also serve as a heuristic warning to a working scientist
("Don't overlook the presence of internal forces opposed to each
other in a given system!"), we would suggest rather that a stronger
interpretation is feasible, for two reasons: 1) the historical origins of
dialectical theory in Hegelian philosophy make it implausible that
the principle is meant merely as a generalization of scientific prac-
tice. Surely, Hegel's dialectic (and the Marxist variant of it) is meant
to disclose the necessary structures and dynamics of human thought
or of reality (or of both). Generalizations are liable to counter-
example and have at best a contingent status, neither of which seem
likely for a basic principle of the (Hegelian-Marxist) dialectic.
2) It would be better to have a view of all three interpretations
(indeed of all six) which gives them a uniform status. McGiU and
Parry's account tends to motivate a view of interpretations #5 and
#6 as also being generalizations from empirical matters. While,
perhaps such generalizations can lead to revisions of one's basic logic
(as some argue today with respect to quantum theory), philosophical
discussion over the past decades (cf. Haack 1974) indicates that it is
by no means obvious that empirical theories can easily give rise to
fundamental alterations in one's logic. Any clash between empirical
practice and classical logic is likely to be interpreted as revealing a
failure in the practice. A view which avoids such problems seems
worth trying out.
192
STUDIES IN PARACONSISTENTLOGIC I
193
NEWTONC.A. DA COSTAand ROBERTG. WOLF
194
STUDIES IN PARACONSISTENTLOGIC I
195
NEWTONC.A. DA COSTAand ROBERTG. WOLF
196
STUDIES IN PARACONSISTENTLOGIC I
197
NEWTON C.A. DA COSTA and ROBERT G. WOLF
198
STUDIES IN PARACONSISTENTLOGIC I
199
NEWTONC.A. DA COSTA and ROBERT G. WOLF
II. Metatheorems of DL
Theorem 1. - schemata and rules of classical positive proposi-
All
tional logic are valid in DL.
P r o o f . - Immediate from AI-A10, which were chosen precisely
so that this would be immediate.
T h e o r e m 2. - In DL, the following schemata (among others)are
not valid:
O) AV7 A
(~) (A2B)~. 7AVB
(iii) A A T A . DB
(iv) 7(AATA)
(v) (A D B) D 7 ( A A 7 B)
(vi) AD(TADB)
(vii) (ADTA) D7 A
(viii) AATA. ~B
(ix) BDAVTA
(x) (TADA) DA
200
STUDIES IN PARACONSISTENTLOGIC I
(ra) -7 A D (A D B)
(xa) AV(TAAA ~
(raa) (AV 7 A)A((A D B ) V ( T A D B)) . D A ~
(xiv) (ADB) D ( T B D 7 A )
(xv) ( 7 B D 7 A ) D(A D B)
(xv)' (AV 7 A. ~ S)V(AA 7A). D 7 A ~
Proof. - It is sufficient to employ the following tables where 2
and 3 are the distinguished values:
A 7 A A~
0 0 0
1 2 2
2 1 2
3 3 0
A B A DB AAB AVB
0 0 2 0 0
1 0 2 1 0
2 0 0 0 2
3 0 0 1 2
0 1 2 1 0
1 1 2 1 1
2 1 1 1 2
3 1 1 1 3
0 2 2 0 2
1 2 2 1 2
2 2 2 2 2
3 2 2 3 2
0 3 2 1 3
1 3 2 1 3
2 3 3 3 2
3 3 3 3 3
Verifying that these matrices actually do the job that we claim they
do, we leave up to the reader.
Before moving on, we should perhaps comment, briefly, on the
significance of the non-theorems: the non-theoremhood of (iii), (vi),
(viii) and (xi) are absolutely essential to the paraconsistent proper-
201
NEWTONC.A. DA COSTA and ROBERT G. WOLF
0 0 1 0 0 0 0 0
1 0 0 1 0 1 0 0
0 1 1 1 0
1 1 1 1 1
Again verification is left to the reader. The significance of theorem
2' has already been discussed.
Theorem 3. - In DL, we have:
(xvi) I- (A ~ B) D A) D A
(xvii) ~-(A ~ .B VC) -- .(A ~ B) V(A ~ C)
(xvi~) ~-(A ~ B) V (B ~ A)
(xix) I-A~ A A T A . D B
(xx) I-A~ A A-IA. D -1B
Proof. - of (xix):
(a) A~ A T A AA F- (A D B) V ( T A D A) A17
But (b) A ~ A 7 A A A, A D B }- B (A3)
and (c) A~ T A D B ~-B A3
Therefore (d) A ~ A 7 A A A I- B Q.E.D. (a), (b), (c)
202
STUDIES IN PARACONSISTENT LOGIC I
(xx) can be proven in exactly the same way (xvi)-(xviii) are left to
the reader.
That (xvi)-(xviii) are theorems shows the nearness to C of D L . D
and V have a classical, non-intuitionistic character, that remains even
when our non-classical -~ is added. The theoremhood of (xix) and
(xx) are steps toward showing that our stability connective does
mean that certain propositions are classically well-behaved; it a well-
behaved proposition A is both true and false, then all hell breaks
loose. In any formalized dialectical theory, only the non-well-
behaved propositions can be self-contradictory. (xix) and (xx) also
provide welcome technical evidence~that o u r - ] is indeed a type of
negation, since for well-behaved formulas, it takes on the properties
of classical negation - as we shall see below.
We now proceed to expand on our claim that well-behaved propo-
sitions obey all the laws of classical thought.
T h e o r e m 5. - In DL, we have:
(xxi) A ~ [--A ~ V T A ~
(xxii) B~ f- (A ~ D B ~ D ((A ~ D 7 B ~ D 7 A ~
(xxiii) A ~ ~- A ~ D (-7 A ~ D B ~
203
NEWTON C.A. DA COSTA and ROBERT G. WOLF
It is perhaps not out of place to point out that definitions 1-3 are
quite standard in discussions of non-classical logic. Ais to be read as
denoting a 'strong' contradiction (and hence is stronger than the si-
milar constant which denotes only a given falsehood). Given
theorem 4 above, we would expect ~ to obey all the laws of classical
logic, giving us the interpretation of C as a subsystem of DL which
we said that we needed. That is does is the point of:
204
STUDIES IN PARACONSISTENTLOGIC I
Theorem 7. - In DL:
(xxiv) ~- (A~B)~((A~B)3--A
(xxv) I-- ~ A D (A D B)
(xxvi) 1- AV~A
Proof. - Immediate, taking into account the definition of ~ and
the postulates of DL.
Corollary. - D,A,vand ~ satisfy all the postdates of classical pro-
positional logic (which is, therefore, in ~a~eertain sense, contained in
DL).
Proof. - Immediate, from a recognition that (xxiv)-(xxvi) are a
sufficient base for C
205
NEWTON C.A. DA COSTA and ROBERT G. WOLF
206
STUDIES IN PARACONSISTENTLOGIC I
logics has made dear, a number of philosophers find the idea of in-
complete and/or inconsistent possible worlds incomprehensible. Our
approach here bypasses those problems and even promises to make
sense out of such locutions as "impossible possible world".
3) Such a linguistically oriented semantics fits nicely into our
suggested pragmatic-Kantian view of the principle of the unity of op-
posites. Such a principle regulates our construction of the concepts
and the theories which we apply to reality. Therefore the semantics
of our dialectical logic should speak directly about the theories we
use and not pretend to characterize a language-independent reality.
While such a conceptualist reading is not necessarily the most
defensible or truest to the dialectical tradition, that it is allowed by
the semantics strikes us as a virtue, rather than a defect.
We begin our presentation of the semantics with some (standard)
definitions:
207
NEWTON C.A. DA COSTA and ROBERT G. WOLF
o o A o
10) A,B~176176 ) ~p;
A ~ E V"~, A ~176E~
11) A~ ~ P = ~ A V - I A , ~ A V ~ 7 A Ef '
12) 7A~ ~ orAATAE [~
Proof. - Similar to the corresponding one o f the classical proposi-
tional calculus (D,A, V a n d ~ have all the properties o f the classical
implication, conjunction, disjunction and negation).
We now proceed to the crucial notion o f a valuation:
Definition 6. - A valuation of DL is a function v:/~ -+(0,1) such
that:
1) v(A D B) = 1 '~ v(A) = 0 or v(B) = 1
2) v(A A B) = 1 "~ v(A) = v(B) = 1
3) v(A VB) = 1 r v(A) = 1 or v(B)= 1
4) v(7(A/x a))= 1 ,~ ~ T A ) = 1 or v(TB)= 1
5) v ( 7 ( A V B)) = 1 '~ v ( T A ) = v ( T B ) = 1
6) v(A ~ = v(B ~) = 1
v((A D B) ~) = v((A A B) ~) = v((-I A) ~ = v ((A V B)~ = 1
7) v(A ~ = 1 ~-v(A ~176= 1
8) v(A~ = 1 ~ v ( 7 7 A D A ) = 1
9) v ( h ) = v ( 7 A) = v(A~ = 0
lO) v(A) ve v ( 7 A) ~ v ( 7 A ~ = 0
We list some properties of valuations.
Theorem 10. - I f v is a valuation, then:
1) v(A) = 1 ~ v ( ~ A ) = 0
2) v(A) = 0
3) v(~'~ = 1
4) v(A ~ = 1 ~ v ( A ) = 1 or v ( 7 A) = 1
5) v(A~) = 1 =~ v ( i ) = 0 or v ( 7 A) = 0
6) v(A ) = v(B ~ = v ( g D B) = v ( g D 7 B) = I =~ v ( 7 A) = I
7) v(g~ = 1 ~ v ( 7 7 A = A ) = 1
8) v(A ~ = 1 ~ v(A V - I A ) = v(~AV~ 7 A) = 1
9) v ( T h ~ = 1 ~ v ( A V 7 A ) = 0 or v ( h A 7 A) = 1
208
STUDIES IN PARACONSISTENT LOGIC I
209
NEWTON C.A. DA COSTA and ROBERT G. WOLF
Corollary 1. ~ ~- A r ~ A
Proof. - Immediate from theorems 11 and 14.
Corollary 2. - ~ A r A
Proof. - Immediate from corollary 1.
We have not spelled out our completeness proof in any detail;
this, we emphasize, is because the details are quite standard, not be-
cause we are hiding anything. The thesis by Alves and the papers by
da Costa and Alves mentioned earlier can be consulted, if there are
any doubts.
We note also:
Theorem 15. - There are inconsistent (7-incomplete) sets o f for-
mulas which have models. A set of formulas has a model iffit is not
trivial.
Proof. - Immediate from theorems 12 and 13 and their corolla-
ties.
Our semantics is a straightforward generalization of that for C
and makes dearer the meanings o f ~- and 7 9 In our semantics, all
propositions are true or false, but not both (which contrasts with the
situation studied in Dunn 1976). See Rescher 1969, pp. 144-148 for
a discussion of the significance of this feature of our semantics.
Nonetheless, a proposition A and its negation, 7 A, may be both
true or both false. If A is well-behaved, then if A is true, then 7 A is
false, and if A is false, then 7 A is true. That 7 should have these
properties is by now, we think, obvious.
As promised, we can get further information about DL from our
semantics. Specifically we can see both how our semantics leads us
to a rejection of double negation and how our use of semantics
designed for paraconsistent logics leads us to deviate from our
original source of inspiration - McGiU and Parry 1948.
Definition 9. - If k is a natural number greater than 0, then 7 k
will be an abbreviation of A proceeded by k occurrences of the sym-
bol 7 .
The following lemmas are not difficult, if one examines our defi-
nition of a valuation.
L e m m a 1. - In DL:
L e m m a 2. - In DL:
(xxxvi) , F 7 kp -- -3 lp, where p is a propositional variable, and k
and 1 are two different natural numbers greater than 0.
Hint. - Our definition of a valuation allows us to define the value
210
STUDIES IN PARACONSISTENTLOGIC I
211
NEWTONC.A. DA COSTA and ROBERT G. WOLF
212
STUDIES IN PARACONSISTENTLOGIC I
213
NEWTON C.A. DA COSTA and ROBERT G. WOLF
214
STUDIES IN PARACONSISTENT LOGIC I
INSTITUTE OF MATHEMATICS
UNIVERSITY OF SAO PAULO
SAO PAULO, S.P.
BRAZIL
NOTES
The authors would like to thank A.I. Arruda and A. Lopafi6, both for the
suggestions and criticisms each has made and for the inspiration their
work has provided. Also we would like to thank E.H. Alves for his
comments and for his own important work on the smbject. Wolf would
215
NEWTON C.A. DA COSTA and ROBERT G. WOLF
also like to thank the National Endowment for the Humanities for fund-
ing him during the period in which this paper was written.
Note that our metalanguage is classical; e.g., A # ~means "A does not
belong to[~''. We could, we think, use DL in our metalanguage, making
our semantics fully dialectical - in much the same way that recent intui-
tionism is concerned to have a fully intuitionistic semantics and that
some recent work in relevant logics is concerned with fully relevant
proofs. But we do not feel that anything vital would be gained at this
stage.
BIBLIOGRAPHY
Aires, Elias Humberto: 1976. Lb~ca e lnconsist~ncia: Urn Estudo dos CMcu-
los Cn. Universidade de Sat Paulo masters thesis.
Anderson, Alan Ross, and Belnap, Nuel D., Jr.: 1975. Entailment." the Logic'
of Relevance and Necessity. Princeton (Princeton University Press).
da Costa, Newton C.A.: 1974. "On the Theory of Inconsistent Formal Sys-
tems", Notre Dame Journal of Formal Logic', vol. 15, pp. 497-510.
da Costa, Newton C.A., and Alves, Elias Humberto: 1976. "Une S6mantique
pour le Calcul C 1", Cornptes Rendus Hebdomadaires des S~ances de
l'Acad~m& des Sciences, s6rie A, vol. 283, pp. 729-731.
da Costa, Newton C.A., and Alves, Elias Humberto: 1977. "A Semantical Ana-
lysis of the Calculi Cn", Notre Dame Journal of Formal Logic', vol. 18,
pp. 621-630.
Dunn, J. Michael: 1976. "A Kripke-style Semantics for R-Mingle Using a Bi-
nary Accessibility Relation", Studia Logica, vol. 35, pp. 163-172.
Gbdel, Kurt: 1932. "Zum lntuitionistischen Aussagenkalkiil", Akadernie clef
ICqssenschaften in lCqen, Mathematisch-naturwissenschaftliche Klasse,
Anzeiger, vol. 69, pp. 65-66. Reprinted in Ergebnisse Eines Mathemati-
schen Kolloquiurns, Heft 4 (1933), p. 40.
Haack, Susan: 1974. Deviant Logic, London (Cambridge University Press).
Hilpinen, Risto: 197+. "Approximate Truth and Truthlikeness".
Fine, Kit: 197, "Vagueness, Truth and Logic", Synthese, vol. 30, pp.
265-300.
Kamp, Hans: 1975. "Two Theories About Adjectives", Formal Semantics of
Natural Language, edited by F_.L Keenan, London (Cambridge University
Press), pp. 123-155.
McGill, V.J., and Parry, William Tuthill: 1948. "The Unity of Opposites: a
Dialectical Principle", Science and Society, voL 12, pp. 418-444.
Meyer, Robert K.; 1974. "New Axiomatics for Relevant Logics. I", Journal o f
Philosophical Logic, vol. 3, pp. 53-86.
216
STUDIES IN PARACONSISTENT LOGIC I
Meyer, Robert IC, and Routley, Richard: 1974. "Classical Relevant Logics II",
Studla Logica, vol. 33, pp. 183-194.
Rasiowa, Helena: 1974. An Algebraic Approach to Non-classical Logics.
Amsterdam (North-Holland Publishing Company) and New York (Ameri-
can Elsevier Publishing Company, Inc.)
Rescher, Nicholas: 1969. Many-valued Logic. New York (McGraw-Hill Book
Company).
Roufley, Richard, and Meyer, Robert K.: 1976. "Dialectical Logic, Classical
Logic, and the Consistency of the World", Studies in Soviet Thought, vol.
16, pp. 1-25.
Routley, Richard, and Meyer, Robert K.: 1978. Relevant Logics and their
Rivals, Department of Philosophy, The Research School of Social
Sciences, The Australian National University, Monograph Series.
Thomason, Richmond H.: 197+. "Supervaluations, the Bald Man, and the Lot-
tery".
Wolf, Robert G.: 1977. "Objectivity in Logic: a Phenomenological Ap-
proach", Interdisciplinary Phenomenology, edited by Don lhde and Ri-
chard M. Zaner, The Hague (Martinus Nijhoff), pp. 169-185.
217