Leibniz's Principle of the Identity of Indiscernibles Author(s): Fred Chernoff Source: The Philosophical Quarterly, Vol. 31, No. 123 (Apr., 1981), pp. 126-138 Published by: Oxford University Press on behalf of the Scots Philosophical Association and the University of St. Andrews Stable URL: http://www.jstor.org/stable/2218719 . Accessed: 24/04/2014 11:38 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Oxford University Press, Scots Philosophical Association, University of St. Andrews are collaborating with JSTOR to digitize, preserve and extend access to The Philosophical Quarterly. http://www.jstor.org This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 126 LEIBNIZ'S PRINCIPLE OF THE IDENTITY OF INDISCERNIBLES BY FRED CHERNOFF There have been numerous attempts to determine what Leibniz's grounds are for his principle of the identity of indiscernibles. In this study I shall question the assumption that there is a single, unique foundation for that principle. By analysing specific arguments, and most importantly, by dis- tinguishing two sorts of question that Leibniz addresses, the supposition of a unique ground turns out to be spurious. We need to separate two types of question: whether God could have brought about states X and Y conjointly; and whether God could have brought about state X or, alternatively, state Y, but not a single world that contains both X and Y. I. INTRODUCTORY Leibniz's final views on a broad range of questions are presented in the Leibniz-Clarke correspondence.1 In trying to excavate the ground of the principle of the identity of indiscernibles [PII], as it appears there, it is usual to restrict oneself entirely to the correspondence. Recourse to other writings is not generally regarded as instructive here. The reason cited is Leibniz's alleged inconsistency on fundamental questions of metaphysics. It is held that, in the correspondence, Leibniz grants the reality of material, spatio-temporal objects in a way denied elsewhere in his writings.2 This paper may provide further reasons for heeding this advice. For if it can be shown that there is more than one version of PII in the correspondence, it follows a fortiori that there is more than one version inherent in the Leibnizian corpus. The question usually raised in connection with the ground of PII is, 'Is it presented as a necessary or as a contingent proposition?'.3 But we must be very careful when formulating the problem in these terms. It is only too easy to equivocate on 'necessary', particularly in connection with Leibniz. There are a number of commonly employed senses of 'necessity', and Leibniz scrupulously distinguishes several of them.4 Some commentators (e.g., Res- cher, Russell, Vinci and perhaps Adams) have argued that PII, in the corres- 1The original edition was edited, and Leibniz's letters were translated, by Clarke. The most recent editions are edited by H. G. Alexander and by Loemker: see list of references below. 2Montgomery Furth discusses this supposed inconsistency, as does Thomas Vinci. 3Among the commentators who pursue this question are G. H. R. Parkinson, Nicholas Rescher, Bertrand Russell and P. F. Strawson. 4In his fifth letter to Clarke, as well as in his Theodicy and "On the Radical Origina- tion of Things", Leibniz distinguishes metaphysical from physical necessity, logical from moral necessity, and absolute from hypothetical necessity. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ'S PRINCIPLE 127 pondence with Clarke, is a necessary proposition, and some (e.g., Parkinson and perhaps Strawson) have argued that it is contingent. Some of those who hold it to be contingent do so because the set of premises from which it is deduced includes propositions (concerning, for example, the nature of the Divine Will) which are not necessary truths. Others (e.g., Vinci) have argued that, since Leibniz views these premises as necessary, PII is presented as a necessary proposition. Thus even when agreement has been reached regard- ing the premises from which Leibniz deduces PII, there has been discord over its own nature. At least some of this disagreement rests upon an equivocation. Leibniz holds that a principle like 'God wills only what is best' is necessary in one sense, but not in another. He believes it to be morally necessary; but its denial is not self-contradictory, so it is not absolutely necessary. We can elude this problem by altering our nomenclature. I shall make use of the distinction between "logical" and "non-logical" propositions for the pur- poses of investigating the ground of PII. "Logical" covers propositions that are true in virtue of our conceptual capabilities and limitations (that is, those whose negations are inconceivable). It is not possible to conceive of an armadillo as having four legs while also having fewer than four legs, nor is it possible to conceive of a set as having many members while also being empty. Truths of logic and set theory are included under the rubric 'logical', which corresponds, in the Leibnizian lexicon, to 'absolutely necessary'. All other propositions, including those Leibniz holds to be morally - but not absolutely - necessary, are non-logical. Hence some claims that PII is necessary assert it to be a logical truth, while other such claims assert it to be non-logical. II. LEIBNIZ'S THIRD LETTER AND THE LOGICAL PRINCIPLE Leibniz professes to provide "many demonstrations to confute the fancy of those who take space to be . . . an absolute being". The first of these demonstrations is presented at L III. 5.5 He offers a negative argument, against the Newtonian conception of space as absolute, and a positive argu- ment, supporting his own relational theory of space. The negative argument is the following: I say then, that if space was an absolute being, there would something happen for which it would be impossible there should be a sufficient reason. Which is against my axiom. And I prove it thus. Space is something absolutely uniform; and, without the things placed in it, one point of space does not absolutely differ in any respect whatsoever from another point of space. Now from hence it follows, (supposing space to be something in itself, besides the order of bodies among themselves,) that 'tis impossible there should be a reason, why God, preserving the same situations of bodies among themselves, should have placed them in space after one certain particular manner, and 5I follow the conventional reference notation: 'LIII. 5' is paragraph five of Leibniz's third letter. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 128 FRED CHERNOFF not otherwise; why every thing was not placed the quite contrary way, for instance, by changing East into West. The negative argument makes explicit use of the principle of sufficient reason [PSR]. The argument can be restated as follows: (N1) Space is absolute, in Newton's sense. (N2) Space is homogeneous: any two points of empty space are the same in every respect. (N3) The universe is ordered in some determinate way. (N4) The present order of the universe is distinct from some contrary order, e.g., the order that would exist if East were instead West. (N5) There is no sufficient reason why God orders the universe in the way that He does and not in some other way. (N6) Something happens for which there is no sufficient reason. .'.(N7) If space is absolute, then PSR is false. The first step is the hypothesis to be considered. The second step is presented as a tenet of both theories, and is not questioned by Clarke, since it is indeed a consequence of Newton's doctrine that Euclidean geometry describes physical space. The third step is the worthy observation that the universe exists and is arranged in some particular way. (N4) is a consequence of (N1), since (N1) implies that there are "absolute directions" of East and West. (N5) is implied by (N4), given Leibniz's conception of "sufficient reason". According to this conception, a will, even God's will, cannot act without a sufficient reason (beyond the mere "will to act"). (N6) is a valid generalization of (N5), given the existential assumption, contained in (N3), that there is a world. The hebdomad is completed by (N7), which follows from (N1) and (N6) by conditionalization. The objection Clarke raises to this argument concerns Leibniz's use, in (N5), of "sufficient reason". Leibniz says that there is "a sufficient reason why things should be so, and not otherwise" (L II. 1). Clarke replies, "But this sufficient reason is oft-times no other, than the mere will of God" (C II. 1). Leibniz abhors this exception to God's perfectly rational nature, and retorts that Clarke "grants me this important principle. . . But he grants it only in words, and in reality denies it" (L III. 2). From here ensues a debate over the nature of God's will. But we can conclude, at this point, that the validity of the negative argument at L III. 5 hinges upon the Divine applic- ability of PSR, that is, upon whether God's will to act can itself constitute a reason for his acting. The positive argument, subjoined to the negative one, is the following: But if space is nothing else, but that order or relation; and is nothing at all without bodies, but the possibility of placing them; then those two states, the one such as it now is, the other supposed to be the quite contrary way, would not at all differ from one another. Their 6The identity of a point of Newtonian space is independent of any matter it may contain. (Cf. Newton's Principia, scholium to definition 8, sects. II-LV, in Alexander.) Therefore any axes and directions can be defined solely in terms of points of absolute space. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ'S PRINCIPLE 129 difference therefore is only to be found in our chimerical supposition of the reality of space in itself. But in truth the one would exactly be the same thing as the other, they being absolutely indiscernible; and consequently there is no room to enquire after a reason of the preference of the one to the other. This argument differs from the negative one in two important respects. First, it is independent of the interpretation of PSR. Its validity is un- affected by whether or not we choose to interpret "sufficient reason" to include God's mere will to act. PSR is used to show only that no conflict with the relational theory arises. And no conflict arises on either Leibniz's or Clarke's interpretation of PSR. Secondly, this argument uses PII. The positive argument can be recast as follows. (P1) Space is relational. (P2) The state of the universe as it is now (SN) and the state "the quite contrary way" (SC) are indiscernible. (P3) If two states are indiscernible, then one is identical with the other (PII). (P4) SN = SC. (P5) It is not the case that God produces the state SN to the exclusion of SC. .. (P6) We cannot infer the denial of PSR (as we could in the negative argument). .(P7) It is not the case that the relational theory implies the denial of PSR. The negative and positive arguments combine to yield Leibniz's con- clusion in the following way. Let 'A' = 'Space is absolute', and 'R' = 'Space is relational'. (01) PSR Axiom (02) A -+ r PSR (N7) ..(03) A (01), (02), MT (04) -(R+ PSR) (P7) (05) A v R Assumption ..(06) R (03), (04), (05) Although this is a plausible representation of the text, the argument is redundant. Given the truth of (05), we can do without (04), the conclusion of the positive argument. Leibniz often expresses his belief in (05). So it is perhaps because (05) is not a truth of logic, nor is anywhere proven by Leibniz, that he includes (04) in the overall argument. But this is not be- yond question. (04), i.e., the positive argument, may be included for rhetor- ical purposes, to lend additional weight to what has already been proven. Or perhaps it is included to show that we need not seek further disjuncts in (05). In any case, Leibniz's firm belief in (05) is clear from such state- ments as "If space and time were anything absolute, that is, if they were any thing else besides a certain order of things . . ." (L IV. 16). The positive argument is at least intended to show that the relational theory does not suffer from the same defect as Leibniz attributes to the This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 130 FRED CHERNOFF Newtonian theory. The conclusion, which I paraphrased as (P7), does not actually follow from (P1)-(P6). It requires qualification. The positive argu- ment does not prove that the relational theory is consistent with PSR. Leibniz has shown that the absolute theory conflicts with PSR, in a specific way. The positive argument reveals only that the relational theory does not conflict with PSR in that particular way. It is still possible that there may be some other point of tension between them. But even this qualified conclusion serves Leibniz's purpose, if that purpose is to demonstrate the superiority, at least in this respect, of Leibniz's theory over its rival. Let us return to the analysis of the positive argument. The feature of the argument that most interests us is (P3), that is, PII. Its use is evident in the sentence, 'But in truth the one would exactly be the same thing as the other, they being absolutely indiscernible'. In order to determine whether or not PII here is a logical principle, let us examine the role it plays in the argument. The proof begins by hypothesizing space to be relational. From this it follows that the "spaces" SN and SC are orders, that is, sets of rela- tions. Let us simplify the analysis by considering Kant's example of a universe comprised of a single hand. Suppose this lone object is a left hand. Space would then consist of the relations rl, r2, r3, . . . constituting the set {rl, r2, r3, .. .}, which I shall call U. Now change East into West and consider a universe consisting of a right hand. Let us call the set of spatial relations of this universe U'. As Kant observed, the points of a left hand are related to one another just as the corresponding points of a right hand are related to one another. For every ri in U (or U') there is an rj such that ri = r, and rj is a member of U' (or U). Therefore, by the axiom of extensionality, U = U'. This justifies the qualified conclusion of the positive argument. God can no more prefer U to the exclusion of U' than He can prefer A to the exclusion of A. Thus the relational theory does not violate PSR in the way the absolute theory does. Indeed, "there is no room to enquire after a reason of the preference of the one to the other", since there is no one in addition to the other. That PII is here a logical principle can be seen by recognizing that Leibniz must deny that it is possible even to conceive that SN = SC. Leibniz often uses the fact that the relational theory implies that space is something ideal, in order to contrast it with the absolute theory. In the positive argu- ment Leibniz postulates that the material world is the same in both SN and SC, except for the spatial relations among material points. Leibniz here uses PII to argue that two ideal entities (points of spatial relations) are identical. But if it is possible to conceive two ideal entities as distinct, then surely they are distinct. So if the positive argument is valid, we must con- strue PII as denying the possibility of conceiving them to be distinct. Thus PII, as it is employed here, is an example of what I have termed a "logical principle".7 7Note that the question addressed by Leibniz and Clarke here is not 'Could God produce a possible world containing both SN and SC?'. It is rather, 'Could God produce This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ' S PRINCIPLE 131 III. LEIBNIZ'S FOURTH LETTER AND THE TWO PRINCIPLES The positive and negative arguments are attacked by Clarke at C II. 2 and C III. 5. He denies that PSR applies to Divine decrees. He maintains that a reversal of the order of three indiscernible bodies constitutes a real change, even if space is "only the mere order of bodies". Likewise, if the earth were instead placed where some distant star is, it would occupy a genuinely different place. And Clarke claims that these cannot be accounted for by the relational theory. Leibniz's reply at L IV. 1-6 makes important use of PII. We can distinguish three arguments. Two, at L IV. 1-4, attempt to prove PII, and one, at L IV 6, uses PII as a premise. The fourth letter begins as follows: 1. In things absolutely indifferent, there is no [foundation for] choice; and consequently no election, nor will; since choice must be founded on some reason, or principle. 2. A mere will without any motive, is a fiction, not only contrary to God's perfection, but also chimerical and contradictory; inconsistent with the definition of the will, and sufficiently confuted in my The- odicy. 3. 'Tis a thing indifferent, to place three bodies, equal and perfectly alike, in any order whatsoever; and consequently they will never be placed in any order, by him who does nothing without wisdom. But then he being the author of things, no such things will be produced by him at all; and consequently there are no such things in nature. 4. There is no such thing as two individuals indiscernible from each other. The first argument, which seeks to prove PII, can be restated as follows: (M1) A will, by definition, requires a motive in order to act. (M2) There can be no motive to order several indiscernible individuals in some way, rather than in some other. (M3) The will of God, though Divine, is nevertheless a will. .'.(M4) God does not order indiscernible individuals to be ordered in any way. .. (M5) God does not produce indiscernible individuals. (M6) God is the creator of all nature. .'.(M7) No indiscernible individuals exist in nature. A second deduction of PII can be abstracted from L IV. 3-4, when we replace (M1)-(M3) above by: (M1') 'Tis a thing indifferent to place two or more indiscernible individuals in any order whatsoever. (M2') Wisdom could not move God to order several indiscernible individuals in some way, rather than in some other. (M3') God does nothing without wisdom. (M4)-(M7) follow as in the above argument. a possible world containing either SN or SC?', that is, 'Could He produce one, to the exclusion of the other?'. And secondly, it is here assumed that PII governs aggregates based on simple substances in the same way that it governs simple substances. Those who deny this assumption face the formidable task of explaining how the aggregates could be subject to different rules of identity. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 132 FRED CHERNOFF The version of PII deduced here is clearly a non-logical principle. It tells us nothing about logic, nor about what it is possible to conceive. It tells us only that, in the actual world, we shall never encounter two indiscern- ible individuals. Furthermore, several premises required in the deductions are not absolutely necessary truths of logic. To see this we should recall that Leibniz assiduously affirms that God is free to actualize any possible world. It is not absolutely necessary that God actualize our own, best possible world. He chooses this world only because He is wholly rational and morally perfect. Thus neither (M3) nor (M3') is a truth of logic (although they are morally necessary). Nor is (M6) a truth of logic. So both arguments contain non-logical propositions as premises. Since PII is here deduced from non-logical premises, Leibniz seems to think it a non-logical principle. At L IV. 6 Leibniz uses PII in the following argument: To suppose two things indiscernible, is to suppose the same thing under two names. And therefore to suppose that the universe could have had at first another position of time and place, than that which it actually had; and yet that all the parts of the universe should have had the same situation among themselves, as that which they actually had; such a supposition, I say, is an impossible fiction. We may ask whether the non-logical version of PII deduced above is the version employed here. In order to answer this, let us distinguish two questions that are addressed by Leibniz and Clarke: (Q1) Could the entire material world have occupied some region of space other than that which it in fact occupies? (Q2) Could the entire material world be moved from one region of space to some other? For the purposes of our inquiry into the ground of PII, the most salient difference between these two questions can be stated in terms of Leibniz's notion of "possible worlds". (Q1) asks about two states that, if distinct, must comprise parts of separate possible worlds. (Q2), on the other hand, asks about two states that, even if distinct, might constitute parts of the same possible world. These two questions are examples of two more general types of question. Leibniz's notion of "compossibility" applies to a set of states if and only if there is some possible world which includes them. We can then characterize the two relevant types of question thus: (Q1') Are the noncompossible, if distinct, states X and Y distinct? (Q2') Are the compossible (even if distinct) states X and Y distinct? It is by recognizing the existence of these two types of question (and that Leibniz undertakes to answer examples of both types) that we can best see the different versions of PII that appear in the correspondence. The argument at L IV. 6 considers the question of whether there could be two specific, distinct states. They are: the material world having a certain position at the time of creation, and the material world, being itself This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ S PRINCIPLE 133 the same, but having some other position at the time of creation. Thus it is (Q1) that is at issue. If the two states are distinct, then they are members of different possible worlds (since any possible world has, by definition, at most one creation state). We cannot then interpret the version of PII used at L IV. 6 as the non-logical one deduced at L IV. 1-4.8 That principle tells us only what we shall or shall not find in the actual world. It cannot relate states of diverse possible worlds to one another. But the present question considers the relationship between two states that, if distinct, are members of different possible worlds. In order to answer this broader question, Leibniz would have to make use of a principle relating states of distinct possible worlds. And indeed that is precisely what he appears to offer, in the argument at L IV. 6. He presents a conceptual principle that denies the possibility of distinguishing, even in thought, the one state from the other: "to suppose two things . . . is to suppose . . .". Therefore, the argument at L IV. 6 both requires, and in fact makes use of, the logical version of PII. Leibniz's fourth letter contains one further argument that interestingly employs PII. It occurs at L IV. 13. To say that God can cause the whole universe to move forward in a right line, or in any other line, without making otherwise any alteration in it; is another chimerical supposition. For, two states indiscernible from each other, are the same state; and consequently, 'tis a change without any change. Besides, there is neither rhyme nor reason in it. But God does nothing without reason; and 'tis impossible there should be any here.9 It appears that (Q2) rather than (Q1) is addressed here. Leibniz is consider- ing the possibility of a single possible world, the actual world, containing two states. They are: the universe occupying some spatial region, and the universe, materially unchanged, occupying some other spatial region, at some other time. The non-logical version of PII is sufficient to generate Leibniz's negative answer to (Q2). Let us then assume for the moment that the non-logical version of PII is intended here. We can detect two arguments in this passage: (R1) The state of the material universe, before it is moved along a straight line, is indiscernible from its state after it has been moved along a straight line. (R2) Two states indiscernible from each other are identical [PII]. .. (R3) The two states are identical. .. (R4) God does not move the material universe along a straight line. 8It might be objected that what is really at issue here is not (Q1), since we should interpret L IV. 6 as considering whether the universe could have had at first a different position, given that it now occupies its present position. It is then a question of whether the world could be moved, in which case (Q2) is at stake. Thus Leibniz can use the non-logical PII, deduced at L IV. 1-4. But this interpretation is not the one closest to the text. Leibniz asks if "the universe could have at first another position . . . than that which it actually had". He thus contrasts two (possibly distinct) initial positions. OI wish only to offer some suggestions for interpreting this very dense passage. It may well be impossible to divine Leibniz's actual intentions. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 134 FRED CHERNOFF And secondly: (S1) There is no reason for God to move the material universe along a straight line. (SI') It is impossible that God should have a reason to move the material universe along a straight line. (S2) God does nothing without reason. .'.(S3) God does not move the material world along a straight line. I assumed for the moment that Leibniz intends a non-logical PII in this passage. But there are at least two considerations that militate against this view, and suggest that he intends a logical principle. First, PII enters into the second argument if anywhere in the justification for the first premise. The non-logical PII tells us that God does not produce two indiscernible individuals in this world. But if we do not invoke the logical PII, then it is possible that the two states (the universe before it is moved along a straight line and as it is after it is moved) are indiscernible in thought, and yet differ solo numero. (Only the logical PII removes this possibility.) In this case God produces neither state, since they are indiscernible.10 But it is obvious that God produced at least one of the two states, since the material world was in fact created. It must then be denied that there are even two states from which God chooses. And this can be accomplished only by recourse to the logical PII. The second point arises from the fact that the two arguments, which I have paraphrased as (R1)-(R4) and (S1)-(S3), are conjoined in a specific way by Leibniz. The first argument makes clear use of PII, "two states indiscernible from each other, are the same state". This appears to be the logical version, for the following reason. The second argument makes refer- ence to PSR. The non-logical version of PII is a consequence of PSR, and thus is not independent of it. Leibniz conjoins the two arguments with the term 'besides', which properly introduces an independent idea. The only way to interpret the second argument, which uses PSR, as independent of the first argument is to interpret the first argument as employing the logical PII. Either of these two considerations is, in itself, far from conclusive proof that L IV. 13 contains only the logical PII. It is even possible that both versions of PII are involved here. We saw that both are at work in L IV. 1-6. So perhaps the logical PII is employed in (R2) of the first argument (R1)- (R4), and the non-logical in the second argument (S1)-(S3). IV. LEIBNIZ'S FIFTH LETTER AND THE NON-LoGICAL PRINCIPLE Leibniz presents a number of arguments that employ PII. We have seen that the validity of some of these requires the logical formulation of PII, while the validity of others does not. Leibniz proffers deductions of lOLeibni firmly endorses this view, e.g., at L IV. 3. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ S PRINCIPLE 135 PII, but they prove only the non-logical version. A cardinal rule of historical interpretation demands that we attribute consistency, wherever possible, to great philosophers. Devotion to this imperative suggests that we try to interpret PII as an unambiguous principle. Since the logical version entails the non-logical version, we might conjecture that Leibniz intends the former. That principle will render valid all of the arguments that employ PII as a premise. But Leibniz's deductions of PII prove only the non-logical version. So this pious suggestion entails that his attempted deductions of PII fail. It does, however, have the gratifying consolation that what is there deduced is not inconsistent with the logical PII. But this consolation, and with it the possibility of interpreting PII as a univocal principle, vanishes when we examine Leibniz's fifth letter, where he repeatedly disavows the logical PII. Leibniz provides another deduction of PII at L V. 21: I infer from that principle [PSR], among other consequences, that there are not in nature two real, absolute beings, indiscernible from each other; because if there were, God and nature would act without reason, in ordering the one otherwise than the other; and that there- fore God does not produce two pieces of matter perfectly equal and alike. This argument proves that no indiscernible individuals exist in nature. Like the proofs in the fourth letter, it demonstrates only the non-logical PII. Leibniz goes on to consider an objection raised by Clarke, and then presents the following perspicuous denial of the logical PII: This supposition of two indiscernibles, such as two pieces of matter perfectly alike, seems indeed to be possible in abstract terms; but it is not consistent with the order of things, nor with the divine wisdom, by which nothing is admitted without reason. The logical PII is denied with equal force at L V. 25: When I deny that there are two drops of water perfectly alike, or any two other bodies indiscernible from each other; I don't say, 'tis absolutely impossible to suppose them .... Leibniz here repudiates a commitment to the logical PII. He does not deny the possibility or conceivability of indiscernible bodies that differ solo numero, as he must in the arguments of the third letter. He denies here only the actuality of indiscernible bodies. This position is taken also at L V. 27: The parts of time or place, considered in themselves, are ideal things; and therefore they perfectly resemble one another like two abstract units. But it is not so with two concrete ones, or with two real times, or two spaces filled up, that is, truly actual. Leibniz again affirms the possibility of individuals that are indiscernible, yet differ in number. It is significant that the denials of the logical PII are accompanied by other arguments that employ PII as a premise. But these arguments require only the weaker, non-logical version. For example, Leib- niz is attempting to subvert the doctrines of atoms and of vacua. The views This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 136 FRED CHERNOFF he opposes assert that atoms and vacua in fact exist. Thus they are (Q2') type questions, rather than (Ql') type questions, in that they refer to the existence of several states within a single possible world. The position advocated by Leibniz on these issues can therefore be advanced by means of the non-logical PII for the same reason as that for which this principle was able to provide him with his answer to (Q2). In his attack upon the existence of atoms, he admits that it is abstractly possible to conceive of several pieces of matter as numerically diverse and yet qualitatively indiscernible. (This is evident, e.g., in the last passage cited.) The theory he is combating maintains that atoms exist in the actual world, not as mere abstract possibilities. The question it answers, as we have noted, is an example of a (Q2') type question. Leibniz can therefore admit the bare possibility of atoms, or the consistency of the concept atom, with- out undercutting any part of his counter-argument. Similarly, his arguments against the existence of vacua do not require the logical PII. He is able to present them using only PSR and its consequences. The most oft-propounded argument against the existence of vacua states, roughly, that any region of space assumed to be void of matter could, without detracting from the perfection of any other matter, have been filled with matter by God.l It is impossible for a void to confer goodness on the universe (or on anything), but it is possible for something which fills the void to confer goodness. It is therefore better to create that region filled with matter than to create it empty. Since God does only what is best, He does not create a vacuum. Leibniz opposes a finite universe with essentially the same argument. If the universe is finite, then it could have been larger. More matter can confer more goodness. So if the universe had been larger, it could have been better. If it could have been better, then it is not the best that is possible. But God acts only for the best. Therefore the universe is not finite. In both argu- ments the premise that God acts for the best is employed. This is not an absolute truth of logic, but rather a morally necessary, non-logical proposi- tion. Thus the conclusions of both arguments should be non-logical. This accords precisely with Leibniz's position on the matter. Absolutely speaking, it appears that God can make the material universe finite in extension; but the contrary appears more agreeable to his wisdom (L V. 30, my italics). Leibniz does not regard the propositions for which he argues, in his attacks on the finite universe, and atoms, as logical truths. He does not need to confine his premises to logical truths. ITe is then at liberty to use non- logical premises such as the weaker version of PII. It is perhaps due to his recognition of this fact that we find in the fifth letter that Leibniz uses only the non-logical PII and disavows the logical PII. He needs only the weaker, less dubitable principle. So he endorses it. He disclaims the stronger llArguments like this occur at L II. 2, L III. 9, L IV. 22-3, L IV P.S. and L V. 43. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions LEIBNIZ'S PRINCIPLE 137 principle, since it is more vulnerable to criticism, and does not, for his present purposes, yield results beyond those obtainable from the weaker principle.l2 V. CONCLUSION There is no single interpretation of PII that has the simultaneous virtues of rendering valid all of the arguments in which PII occurs, and of remaining consistent with Leibniz's repeated remarks, in the fifth letter, on the conceiv- ability of indiscernible individuals. Leibniz employs two logically distinct versions of PII. The third letter contains arguments, at L III. 5 and L III. 6, demonstrating that supposed alternative places and times of creation are, on the relational theory, indiscernible from one another and therefore iden- tical. The alternatives are noncompossible, if they are distinct. Thus the questions at stake are of type (Q1'). The non-logical PII cannot relate such states to one another, and hence cannot be used to manufacture Leibniz's conclusions at L III. 5 and L III. 6. The fourth letter appears to use both the logical and non-logical version of PII. And in the fifth letter Leibniz uses the non-logical PII, while he rejects the logical PII. Any attempt to interpret PII, as it appears in the correspondence with Clarke, as a unique principle that accords with all of Leibniz's relevant remarks is doomed to failure. The type of interpretation we have ruled out may be termed "Athenian", after Castafieda. An Athenian interpretation presumes that the work of a philosopher sprang fully developed from its author, as opposed to a Darwinian interpretation, which allows that changes may occur when conflicting views battle with one another for domination, and the fitter survive. The situation we have encountered in the correspondence vociferously demands a Darwinian approach. The third letter contains only the logical PII. The fourth letter appears to contain both, while only the non-logical version is proven. And the fifth letter uses only the non-logical version, while explicitly denying the logical version. The Darwinian standpoint permits us to suppose that Leibniz intended only one version of PII. As his ideas were challenged by Newton and Clarke, he scrutinized his position more closely, and as the debate came to focus on somewhat different issues, the need for a broader principle diminished. For these and other reasons PII is transfigured from a logical into a non- logical principle. The struggle between the two competing versions of PII appears to culminate in the fourth letter. And the non-logical version emerges victorious in the fifth letter. In a discussion of a different issue, Castanieda states that "The Darwinian approach is particularly necessary in the case a2It might be argued that the shift from the logical to the non-logical PII could be explained away by appeal to the distinction between two entities' sharing all their properties and their sharing only their non-relational properties. But there appears to be no evidence in the correspondence that Leibniz viewed this as the basis of the shift in his use of PII. And in earlier papers (such as "First Truths", Loemker I, pp. 411-7) he seems to endorse both the relational and non-relational applications of PII. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions 138 FRED CHERNOFF of Leibniz".13 The foregoing results substantiate this claim. In considering the ground of PII I have suggested that an Athenian interpretation of that principle, not only as it appears throughout the works of Leibniz, but even as it appears in his last philosophical correspondence, is not possible. Yale University REFERENCES Robert M. Adams, "Leibniz's Theory of Contingency", Rice University Studies, 63 (1977), 1-41. H. G. Alexander (ed.), The Leibniz-Clarke Correspondence (New York, 1970). Hector-Neri Castafieda, "Leibniz's Concepts and their Coincidence Salva Veritate", Nois, 8 (1974), 381-98. Samuel Clarke (tr. & ed.), A Collection of Papers which passed between the late Learned Mr Leibniz and Dr Clarke in the years 1715 and 1716 relating to the Principles of Natural Philosophy and Religion (London, 1717). Montgomery Furth, "Monadology", Philosophical Review, 76 (1967), 169-200. Hid6 Ishiguro, Leibniz's Philosophy of Logic and Language (London, 1972). Leroy Loemker, Gottfried Wilhelm Leibniz, Philosophical Papers and Letters (Chicago, 1956), 2 Vols. G. H. R. Parkinson, Logic and Reality in Leibniz's Metaphysics (Oxford, 1965). Nicholas Rescher, The Philosophy of Leibniz (Englewood Cliffs, 1967). Bertrand Russell, A Critical Exposition of the Philosophy of Leibniz (Cambridge, 1900). P. F. Strawson, Individuals (London, 1965). Thomas Vinci, "What is the Ground for the Principle of the Identity of Indiscernibles in Leibniz's Correspondence with Clarke?", Journal of the History of Philosophy, 12 (1974), 195-201. 10p. cit., p. 383. This content downloaded from 147.91.1.45 on Thu, 24 Apr 2014 11:38:05 AM All use subject to JSTOR Terms and Conditions