Logique ck Analyse 161-162-163 (1998), 57-65

THE OF LOGICAL RELATIVISM

Gert-Jan C. LOKHORST

Abstract We explicate the thesis of logical relativism (people of different cul- tures may have different ) in logical terms. Our illustrations come from the field of .

1. Introduction

Logical relativism is the claim that

People of different cultures may have specifically different logics (for example, [there may bel a peculiarly Chinese logic distinct from West- ern logics). [16]

Logical relativism was rather popular in anthropological circles before the Second World War [11, 14, 151. It then went out of fashion 171. The thesis is currently undergoing a revival in the so-called Strong Programme in the sociology of knowledge [2, I n Cultural anthropologists usually maintain that there is no evidence that there exist cultures which adhere to different logics than we do [23, 241. But I find this a strange claim. For one thing, even in my own country there is a subculture of people who try to adhere to intuistionistic logic rather than classical logic. For another, there are, as far as I know, no anthropolo- gists, logicians or sociologists of knowledge who have ever tried to formu- late the thesis of logical relativism in a truly precise way. As long as the latter has not been done, it does not make much sense to say that there is no empirical evidence for this thesis. In the following, I shall try to make the thesis more precise, and I will use some familiar logical tools to do so. All systems which I will present for illustrative purposes come from the field of paraconsistent logic. 58 GERT-JAN C. LOKHORST

2. Standard Doxastic Logic

When talking about the logical properties of other peoples' opinions, it is natural to turn one's attention to doxastic logic, the logic of [12, 191. It is, however, impossible to make sense of the relativists' claim in terms of standard doxastic logic. Standard doxastic logic is based on modal system K. This system has, amongst others, the following rules. C stands for classical propositional logic.

RI F -R2 F c çcR 3c 1 —F- › Reading D p as "the agent that ço," these rules imply (1) that all FcK agents believe all classical tautologies, (2) that agents' beliefs are closed tDp unde-r C-provable consequence, and (3) that if an agent believes that (p, he P also 1qbelieves all which are C-provably equivalent with q). Thus(-; st an dard doxastic logic portrays all doxastic agents as followers of classD1ical logic. When one wants to deal with logical relativism, one has to adopP1t a less parochial outlook. Romane Clark has written that belief ascription is —1 q >mainly a matter of keeping the references and concepts of those of us Diwho are scribes, recording the occurrences of psychical happenings, P•distinct from those of the agents to whom we ascribe mental events. [6] ; The logical relativist seems committed to the claim that belief ascription involves keeping the agents' and our own logics distinct as well. Just as x y —) (0Fx [ i F y ) is not an acceptable principle of doxastic logic, so 1-c- tli o t p is, from the relativist's perspective, not acceptable either.

3. Adhering to a Particular Logical System

What does it mean to say that somebody adheres to a particular logical system? In the literature about logical relativism, this expression is used in at least two quite different senses. First, it is sometimes said that people may adhere to different logics in the sense of embracing different sets of logical truths (see, e.g., [241). Second, it is sometimes said that people may adhere to different logics in the sense of following different logical rules THE LOGIC OF LOGICAL RELATIVISM 59

(see, e.g., [13]). The first conception of adherence to a logic seems related to R 1 above, whereas the second seems closer to R2 and R3. We accordingly propose the following three definitions:

Def 1 T h e agent adheresi to logical system X according to logical system iff Fx h y 0 c p .

Def 2 T h e agent adheres2 to logical system X according to logical system iff I - Dexf 3 T h e agent adheres3 to logical system X according to logical system tiff t t o ' ) P1 y1The1s e s4ens es of "adherence to a logic" should be carefully distinguished from each other, as the following example makes vivid. -D O .Let CC be the complement of the , i.e., the set of all- non-theorems of C. This system has been axiomatized as follows ([4] see- also [3, 25, 26]). ' D A l p (p atomic) P A2 — p —› p (p atomic) • R la (pip --> çc (p atomic, p does not occur in t Rib i —› ( p atomic, p does not occur in tp) R2 o(P )t i l l (io —› (cP —) R3 till(X ( P ) R4 cP —) 0 R5 1 0 R6 —( c› tPy —11 ( I ) —› tp R7 cP ——› (0 —( › x)/ ( c P —› R8 cp 4S , P— > —1(cp — › S, where S is of the form S = Si or SX S i —) (S2 —* (S n -1 S n )— ), with S, = pi or Si = —1 t Pk for i k , and p E lPi, P2, P n ) for all p which occur in Pi) (P —› qi• , —P i CC is perfectly› unsound and completely antitautological in the sense that kcc (PM' 10c (p141. It will be clear that CC is paraconsistent (it contains all logical falsehoods but no logical truths) and non-monotonic. Now define C-i-ECC as follows.

• Axioms: All classical tautologies. • Rules: I. hC-FOCC DP; 60 GERT-JAN C. LOKHORST

2 .3. Mo d u s Ponens. I It will "be clear that an agent whose beliefs are described by this system does noCt adhere to classical logic in the sense of Def. I. Nor does he adhere to clascsi cal logic in the sense of Def. 2, for we have Fc I - 4h C A-LTIC4C Elb -I- -u4 Dt I . Then agoent tdoes however adhere to classical logic in the sense of Def. 3. For suppose that I-c (1, I l l . It follows that Fc (p iff hc 0 0, whence I-cc cp iff Fcc 0. So either (1) Fcc tp and Foc 0 or (2) l F a,nd k c c 0. In the first case, we have Fc+L cc El cp and F c C +'w hecncce 1 -( c + EcP+asiec, cwe haEve Fc -i-Dtcc p- - , lAEE c-kcp M a n d 1aLn y- tcaspce, QED. C0A It+ is cO n otC difCficu lt to contrive systems in which an agent adheres to a -aparticularI I —I lo gic in the sense of one of the other definitions. One may for example use the logic of awareness discussed in 1121, §9.5. Int i 1 l Dda nDo , P 4whd. Adhh erencee to Differentn Logical Systems ceh e; l W-n hat cdoes it mean to say that two agents adhere to different logics? We pCceropos e the following three definitions, corresponding to Defs. 1-3 above. +Oe1A° m-EeansI thatC agenCt i believes that tp. Ic -D+el 4 Agents i and k adhere C1tEe mt Xoc i ff dit isi fnof t ether ceasne tht a t Hx Eitp <=> F x Okcp. +lc o g i c s LaDef c5 Acgenots i randd k adi here2 to different logics according to logical sys- intWem Xg iff it is not the case that Ex LAI) < = > x Mop -) Ea /. ct) o Def 6 Agents i and k adhere3 to different logics according to logical sys- cl o g i c t-em X iff it is not the case that Ex Elicp D I P i=> Ex Lktp Da- l TPs-he followyi ng systesms illustrate the notion of several agents each follow- =-iEng his own lo gic in the senses of all three definitions. They are simplified I0versions of the systems discussed in 1171. l; Each system Ti + (Ecek)o,k,„, 0 i < to, is defined as follows. li qn • Axioms : A ll axioms of ( it e ik,n oiw.ne p.ar,ac onsistent hierarchy [8], plus Lk(tp —› — › (Elkg) —› for all k, 0 k < co. .h t h e i - Se t h os s y s t e m we i n ec D a ho C o s t an a ' s vd w e l e l - F c + D c C 1 1 1 ( 1 ) ' L i t ' i n THE LOGIC OF LOGICAL RELATIVISM 61

• Rules: and 1-(C k F - D o p , It is of bvioous thrat in each system T i + (Ere, 0 0adheare to differentl l logics in the senses of Defs. 4-6. < Actually,k th, e agen ts adhere to different logics in an even stronger sense of the word. Following [21], one may call a theory of X iff F contains all k < 0 axioms of X and is closed under conjunction and Modus Ponens. One may , , ka l l a g e n t s kthen ob,serve th at eaich agennt k ad, heres to c-Ck in the sense that for all theo- ries

We may explicate this claim in terms of the semantics of the systems we have described. We do not have to refer to non-classical worlds when giving a semantical account of C+OCC. A variation on the usual neighborhood semantics of classical modal systems (see, e.g., 1.51) suffices. Fo r let us define a C+FICC-model as a structure aR = 14 7N W ( ,3fo l)Clo9w,in g Ntwo, co nd1itio,ns /(N,B: 41 df IN' E W : V(tp, w) = 1)): (1) N(w) = w5I l e h3 Veciassricai e1 / W(b(ecWhPa)ve ,s lIikel th;e valuation i functions o f classical logic; on the other hand, if a((p 2 ) C C I . It will be clear that F sn,i f e t - , - Cd=( p 0 -In1 c1on1trCasCt t o the semantics of C+ECC, those of 'Ci + (L re k)0 , k CVDi s 4rhe othFemr h anFd, if cp E lit lf , then V(tp, w) = min{ v(p, w') : w • E Wi and w1,DRw'). H /an cp means that 1/((p, w ) . 1 for all w E W. It can be proven that 0'h4cC i + (E'Ck)o -k

6. Adjacent Territory

Without going so far as to claim that everything is relative, we want to point out that the ideas we have presented are not only relevant in connec- tion with doxastic logic. They may also be applied to, for instance, deontic logic (different cultures may not only have different norms, e.g., in the way outlined in [18]: they may also have different ways of judging adherence to these norms), truth in fiction (one should not judge the works of, fo r example, intuitionists by our —classical or, as the case may be, paraconsis- tent standards) and alethic (the concepts of necessity and possibility are logic-relative). Some of the details may be found in [17]; we leave the rest to the hopefully imaginative reader.

Department of Philosophy, Erasmus University Rotterdam email: lokhorstOffwb.eurnt

REFERENCES

[1] A . I. Arruda. On the imaginary logic o f N. A. Vasilév. In A. 1. Arruda, N. C. A. da Costa, and R. B. Chuaqui, editors, Non-Classi- cal Logics. Model Theory, and Computability. North-Holland, Am- sterdam, 1977. [2] D . Bloor. Knowledge and Social linagery. Routledge and Kegan Paul, London, 1976. [3] P . Bonatti and A. C. Varzi. On the meaning of complementary sys- tems. In Tenth International Congress o f Logic, Methodology and Philosophy o f Science: Volume o f Abstracts, page 122. Florence, 1995. [4] X . Caicedo. A for the non-theorems of the proposi- tional calculus. Notre Daine Journal of Formal Logic, 19:147-151, 1978. [5] B . F. Chellas. Modal Logic: An Introduction. Cambridge University Press, Cambridge, 1980. [6] R . Clark. Old foundations for a logic of perception. Synthese, 33:75– 99,1976. [7] M . Cole and S. Scribner. Culture and Thought: A Psychological Introduction. Wiley. New York, 1974. [8] N . C. A. da Costa. On the theory o f inconsistent formal systems. Notre Dame Journal of Formal Logic, 15:497-510,1974. [9] N . C. A. da Costa and E. H. Alves. A semantical analysis of the cal- culi C n . N o t r e D a i n e J o u r n a l o f F o r m a l L o g i c , 1 8 : 6 2 1 - 6 3 0 , 1 9 7 7 . 64 GERT-JAN C. LOKHORST

HO] M . E. de Montaigne. Apologie de Raimond Sebond. In Essais Simon Mil langes, Bordeaux, 1580. [ I I ] E . E. Evans-Pritchard. Witchcraft. Oracles and Magic Among the Azande. Clarendon Press, Oxford, 1937. [12] R . Fagin, J. Y. Halpern, Y. Moses, and M. Y. Vardi. Reasoning about Knowledge. MIT Press, Cambridge, Mass., 1995. [13] R . C. Jennings. Zande logic and Western logic. British Journal for the Philosophy of Science, 40:275-285,1989. [14] L . Lévy-Bruhl. Les fonctions mentales dans les sociétés inférieures. Alcan, Paris, 1910. [151 L . Lévy-Bruhl. La mentalité primitive. Alcan, Paris, 1922. [16] L . Lévy-Bruhl. Les carnets. Presses universitaires de France, Paris, 1949. [17] G . J. C. Lokhorst. Multiply modal extensions of Da Costa's C„, () n 5_ co, logical relativism, and the imaginary. The Journal of Non- Classical Logic. 5(2):7-22,1989. [18] K . Mengel - Logic of Ethics (1934). Reidel, Dordrecht, 1974. .[1 9] J.-J. C. Meyer and W. van der Hoek. Epistemic Logic for AI and M o r Caoml iptutyer Science. Cambridge University Press, Cambridge, 1995. ,[2 0] L . Z. Puga and N. C. A. da Costa. On the imaginary logic of N. A. Vasilév. Zeitschrift for Mathematische Logik und Grundlagen der D e Mcathi ems atik, 34:205— i o n , [21] P.. K. Schotch and R. E. Jennings. Inference and necessity. Journal of a 2Phn1ilo1so,p1hic9al8 Lo8gi.c, 9:327-340,1980. [d22] D . Sperber. Apparently irrational beliefs. In M. Hollis and S. Lukes, S eoditors, Rationality and Relativism. Basil Blackwell, Oxford, 1982. [c23] Ti . Triaplett. Azande logic versus Western logic? British Journal for l the Philosophy of Science, 39:361-366,1988. [O24] Tr . Tgriplett. Is there anthropological evidence that logic is culturally a nrelaitive? Remarks on Bloor, Jennings, and Evans-Pritchard. British z aJournalt i for the Philosophy of Science, 45:749-760,1994. [o25] An . C.: Va rzi. Complementary sentential logics. Bulletin of the Section of Logic, 19:112-116.1990. T [26] A . C. Varzi. Complementary logics for classical propositional lan- o guages. Kriterion, 4:20-24,1992. w [27] N . A. Vasilév. Voobrazhaemaya (nearistoteleva) logika. Zhu rnal a Minister,siva Narodnogo Pros veshcheniya, 40:207-246, August r 1912. Reprinted in [28]. [d28] N . A . Vasilév. Voobrazhaemaya logika: lzbrannye trudy. lzdatelstvo a Naukya, Moskva, 1989. THE LOGIC OF LOGICAL RELATIVISM 65

[29] E . Zarnecka-Bialy. Truth and logic in Vasilév's imaginary worlds. In Proceedings o f the Bulgarian -Polish Symposium on Logic, pages 11-19. Bulg. Akad• Nauk, Sofia, 1985.