Brody 1973 Why Settle for Anything Less Than Good Old-Fashioned Aristotelian Essential Ism

8/6/2019 Brody 1973 Why Settle for Anything Less Than Good Old-Fashioned Aristotelian Essential Ism

1/15

WhySettle for AnythingLess than GoodOld-FashionedAristotelian EssentialismBARUCHA. BRODY

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

I should like in this paper to explain and defend a theory ofessentialism that I have briefly set out elsewhere.' In particular,I would like to show that (a) it is based upon a simple distinctionthat we can often employ without any difficulty, (b) it is open tonone of the standardobjectionsto essentialism,and(c) it recognizesas essential significant properties of particular physical objects,most importantly, the ones that correspond to the intuitivelyappropriateAristoteliansecondarysubstances. If all of this can bedone, then there will be no reason to offer, as recent defendersofessentialismlike KaplanandPlantingahave, some de dicto reinter-pretation of essentialist claims. This is fortunate, for the theoryin question does not seem to lend itself to such reinterpretations.In short, then, I will be arguing for an unabashed traditionaltheory of Aristotelian essentialism.

IOn my account, all essentialist claims are based upon a

distinction between types of changes first drawn by Aristotle, adistinctionbetween alterations,on the one hand, and comings-to-be and passings-away(which we will refer to from now on assubstantialchanges),on the other hand. He set out that distinctionas follows:Since, then, we must distinguish (a) the substratumand (b) the prop-erty whose nature it is to be predicatedof the substratum;and sincechangeof each of these occurs;there is alterationwhenthe substratumis perceptible and persists, but changed in its own properties, theproperties in question being opposed to one another either ascontraries or as intermediaries... But when nothing perceptiblepersists in its identity as a substratum,and the thing changes as a

351


2/15

352 NOUSwhole... such an occurence s no longer alteration. t is a coming-to-beof the one substance and a passing-awayof the other... . If, however,in such cases, any property persists in the thing that has come-to-bethe same way as it was in the thing which has passed-away, .. thesecond thing, into which the first changes, must not be a propertyofthis persistent identical something. Otherwise, the change will bealteration. [2], 319b 7-24)

In other words, if, before the change, there was an object 01 thathad a property P1, then the change is an alteration if 01 continuesto exist after the change but now has a property P2 whose possessionis incompatible with the possession of P1, while the change is asubstantial change if, after its occurence, 01 no longer exists even ifthere is some 02 (o o1) which has P1 .There are several points to be noted about this distinction.

(1) This theory of substantial change is perfectly compatiblewith the view that there is something that persists through allchanges. To begin with, some of the properties of the old objectcan persist in the new one. More importantly, it is perfectly possiblefor 01 and 02 to be composed out of the same thing or things. Theycould, for example, be composed out of the same particles of matter.All that is necessary is that these permanent entities are not theonly subjects of predication, that 01 and 02 also be subjects ofpredication. This is an extremely important point. After all, thereis a very attractive view that a goal of science is to discover someunderlying permanencies that persist through all changes, and somehave suggested that the discovery of these permanencies willshow that nothing ever stops existing or begins existing, that allthat happens is that these permanencies are altered. This suggestionis mistaken. As we have just seen, the Aristotelian distinction isnot challenged by any such discoveries. Interestingly enough,Aristotle seems to have been aware of this point. For his ownreasons, he had adopted (in [1], Book I, Chapter 7 and [2], Book II,Chapter 1) the very strong claim that there must be somethingthat persists through any change, something other than ordinarysubstances which do not persist through substantial change. This iswhy he postulated the existence of matter.

Matter,in the most propersense of the term, is to be identified withthe substratumof coming-to-beand passing-away. [2], 2-4)But he saw that, despite all of this, he could maintain the distinctionbetween substantial change and alteration.


3/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 353(2) Although it is sometimes supposed that it is a matter ofconvention whether or not a given change is an alteration or asubstantial change, i.e., that it is a matter of convention whethera given 01 persists through a given change, this seems to be amistake. Is it, for example, a matter of convention whether wecontinue to exist after our death ?And if so, why haven't we insuredour immortality a long time ago ? More seriously, it seems obviouslytrue that trees stop existing when they are burnt down, but notwhen they lose a branch, and that this was true a long time beforethere were people who had conventions about the identity of trees.So our distinction cannot be based upon conventions. To be sure,there are undoubtedly many cases in which we have to decidewhether what has occurred is an alteration or a substantial change,and this has perhaps led some people to suppose that the wholedistinction is merely a matter of convention. That is, however,a weak argument; it could, after all, be applied equally to all dis-tinctions about which there are borderline cases, viz., to alldistinctions. The existence of decisions merely indicates that thereare borderline cases, not that the distinction is based upon con-vention.(3) It is also sometimes supposed that whether a given changeis an alteration or a substantial change depends upon how we

classify the object involved when we refer to it. Thus, if I describemy red desk by referring to it by use of 'my red desk', therebyclassifying it as a red object, then a change in its color would con-stitute a substantial change. It would not, however, if I merelyclassify it as a desk by referring to it by use of 'my desk'. Thissupposition also seems to be in error. To begin with, it is open toall of the types of objections raised against the view that the distinc-tion is conventional. Moreover, there seems to be nothing wrongwith saying 'my red desk could someday be painted green withoutits stopping to exist because it is painted green'. There is nothingwrong with saying this because one is referring here, both whenone uses 'my red desk' and when one uses 'it', to the same substance,one that can persist through changes in its color. So red desks,even when referred to in such a way as involves their color, can losethat color and still continue to exist, and the supposition we areconsidering is straightforwardly false.(4) This distinction is applicable to changes involving bothconcrete and abstract objects. In the case of abstract objects, how-ever, something very strange happens. While one can imagine2alterations involving abstract objects-four used to be my favorite


4/15

354 NOUSnumber, but it isn't any more-it seems impossible for there to be asubstantial change involving one. This is due to the fact-at least,to what we all assume, for reasons that are unclear, is a fact-thatabstract objects exist necessarily and cannot go out of existenceas they would if they were involved in substantial changes.(5) This distinction is one that we are all quite familiar withand can use, quite comfortably, in many cases in which questionsof identity and continued existence are of importance. One notableexample of its use occurs when we think about human death. Whatmakes us treat it so differently from other changes involvinghuman beings is precisely that it, but not the other changes, isthe going-out-of-existence-of, and not merely an alteration in,the person in question. This distinction is also used, of course,in more prosaic cases. If your car is insured, then lots of thingscan happen to your car, and it will still be insured (unless, of course,the insurance is formally cancelled). It can be painted, you canchange the engine, etc. But if it is entirely flattened and made intosheet metal from which a car is made, then your insurance wouldnot cover the new car since the new car is not identical with yourold covered car. The changes that the old car underwent, unlikethe paint job and the engine job, were substantial changes and notmere alterations.Given this distinction, we are now in a position to introducea very simple theory of essentialism. We shall say that an object 01has a property P1 essentially just in case 01 has P1 and would go outof existence if it lost it: just in case the loss of it would involve asubstantial change. We shall say that an object 01 has a property P1accidentally just in case 01 has P1 but could lose it without goingout of existence: just in case the loss of it would involve a merealteration.In light of what we have seen so far about our distinctionbetween alterations and substantial changes, we can immediatelysay the following about the distinction between essential andaccidental properties: it is a distinction that applies equally toconcrete and abstract objects; it is based upon neither our con-ventions nor our way of classifying objects; and it is one that we canoften operate with without any difficulty, although there willcertainly be many borderline cases in which we will have to decidewhether a property is essential or accidental.It is important to note that it does not follow from the abovedefinition of 'essentially' that if 01 has a property P1 essentially,then there is some name n, which either necessarily names 01 or is


5/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 355a proper name of 01 and which is such that _P1n1] necessarilyexpresses a truth. Since the existence of such a name is requiredfor any of the presently suggested de dicto reinterpretations ofessentialist claims,3 it must be admitted that they cannot be usedto interpret our claims to those who only admit de dicto modalclaims. But, I submit, if our program is successful, there will beno reason to be disturbed by this lack of a de dicto reinterpretation.

IIThree standard objections have been raised against essen-tialism, one having to do with referential opacity, one having to dowith the different classes to which objects belong, and one havingto do with the problem of identity through possible worlds. It isnow well known that the first two of these objections are basedupon a confusion,4 so I will just briefly explain what is wrongwith them and spend most of my time on the third problem, the onewhich has not yet been dealt with satisfactorily and which our theoryof essentialism solves very neatly.As far as referential opacity is concerned, the objection runsas follows: consider the claims that

(1) Nine is essentially greater than sevenand

(2) The number that I am thinking about now is essentiallygreater than seven,

and suppose that I am now thinking about nine. Then, these claimsshould both be true or both be false since they both say of the sameobject that it has the same property essentially. But, so the argumentgoes, (1) is true and (2) is false. After all, I could have been thinkingabout five, and then the number that I would be thinking aboutnow would be less than seven. So there is something wrong withessentialist claims, and they should be rejected. The answerto this objection is very simple: (2) is true, just as (1) is, for thenumber that I am thinking about now could not be altered so as tobe less than seven. The reason why one thinks that (2) is false isthat I could have been thinking about some other number which isless than seven, but all that that shows is'that the very differentclaim that

(2') 'The number that I am thinking about now is greaterthan seven' is necessarily true


6/15

356 NOUSis false. As long as you do not reinterpret (2) so that it means (2'),and no essentialist should reinterpret his essentialist claims in thisde dicto fashion, there is no problem with referential opacity.A similar error is committed by the objection from the factthat objects belong to different classes. It runs as follows: consider,as an example, the claims that

(3) All cyclists are essentially two-legged but not essentiallyrationaland that

(4) All mathematicians are essentially rational but notessentially two-legged,and consider Joe who is both a mathematician and two-legged.It seems to follow from the above claims that Joe both is and is notessentially two-legged and essentially rational. So our essentialistclaims have led us to a contradiction, and essentialism shouldtherefore be rejected. The answer to this objection is once morevery simple: both (3) and (4) are false since cyclists are not essen-tially two-legged (a cyclist can lose a leg and still continue toexist, although he cannot cycle anymore), and mathematiciansare not essentially rational (a mathematician can stop being rationaland still continue to exist, although he cannot then do any moremathematics.) The only reason why one might think that (3) and(4) are true is because one is confusing them with the true claimsthat

(3') 'All cyclists are two-legged' is necessarily true while'all cyclists are rational' is not,and

(4') 'All mathematicians are rational' is necessarily truewhile 'all mathematicians are two-legged' is not.So long as one does not reinterpret (3) and (4) so as to mean (3')and (4'), and no essentialist should interpret essentialist claimsin this de dicto fashion, there is no problem about objects like Joewho belong to more than one class.We come then to the much more serious problem of identitythrough possible worlds. What has essentialism to do with identitythrough possible worlds, and why is there a problem about identitythrough possible worlds? To deal with the first question first, the


7/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 357believer in essentialism seems to believe that objects have propertiesnecessarily, that there are properties which are not merely had byan object in the actual world but which would be had by it in allpossible worlds (or, at least, all in which it exists). And this seemsequivalent to the idea that it is had by it in the actual world and bythe object identical to it in any given possible world (or, at least,in any world in which there is an object identical to it). So, it wouldappear, essentialist claims are equivalent to claims involving theconcept of identity through possible worlds and are thereforemeaningful only if that notion is meaningful. As far as the secondquestion is concerned, it is easy to see that the notion of identitythrough possible worlds cannot be the same as the ordinary notionof identity. Consider, for example, two actual physical objectsO1and 02 . It is thought to be the case that they are identical just incase they are spatio-temporally continuous (and, perhaps, havecertain properties in common).5 But this cannot be so if 01 is anactual object and 02 exists in some possible world. After all, itseems as though an object could be in a possible world in somevery different place than it is in the actual world, so we haveto allow for the possibility that 01 is identical with 02 even though01 is in P1 at t1 while 02 is in a very different place P2 at t1 and is nottherefore spatio-temporally continuous with 01 . The identity ofan actual physical object with a possible physical object does nottherefore involve any requirement of spatio-temporal continuity,so we must be employing a different conception of identity whenwe talk about their identity. And similar arguments can be con-structed to show that the identity of actual entities that are notphysical objects cannot be the same type of identity as the identityof one of those actual entities with a corresponding possible entity.So there is a problem about identity through possible worlds.It is often thought6 that the problem of identity throughpossible worlds is a problem about how we can tell, of someobject in a possible world, which, if any, object in the actual worldit is identical with, i.e., that the problem of identity throughpossible worlds is an epistemological problem. That is not thereal problem. After all, even if one could not tell in many (or evenall) cases, it would not follow, unless one was using some verysuspicious verificationist assumption, that there is somethingwrong with the concept of identity through possible worlds. Thereal problem is that it is unclear as to what it means for an actualentity el to be identical with an entity e2 that exists in a possibleworld, and it is therefore unclear as to what any claim, e.g., an


8/15

358 NOUSessentialist claim, that involves this concept of identity throughpossible worlds means.There are two accounts of identity through possible worldsthat should be noted, one due to Chisholm (in [10]) and one dueto David Lewis (in [13]). Chisholm has proposed that, given theassumption that there is, for every object, a set of properties whichit has in all possible worlds in which it exists but which is had byno other object in any possible world, one can say that 02 in a pos-sible world is identical with 01 in the actual world just in case theyboth have that set of properties in common. Now there certainlyare objects, like numbers and propositions, for which this assump-tion is plausible. Four is the only object in any world, actual orpossible, which is the sum of two and two, and it is the sum of twoand two in all worlds in which it exists. The proposition that fouris the sum of two and two is the only object in any world, actual orpossible, that is the contradictory of the proposition that four is notthe sum of two and two, and the former proposition is the contra-dictory of the latter proposition in all worlds in which it exists.But for objects like chairs and people, there is no particularreason to suppose that the assumption is true and that Chisholm'sdefinition of identity through possible worlds will work. Indeed,the assumption (and therefore Chisholm's definition) seemsimplausible for such objects. If the set only contains propertiesthat can be had by more than one object in a given world, then onecould easily imagine a world in which two different objects, differentbecause they are in different places at the same time, have all of theirproperties in common. And if the set contains some propertiesthat can be had by only one object in a given world, then it seemsunlikely that the object must have that property in all worlds inwhich it exists. Indeed, it is difficult to think of any such property.Lewis's proposal is quite different. He would drop all talkabout identity through possible worlds and would say rather thanan object has a property essentially just in case it has it in the actualworld and its counterparts in all possible worlds in which it has onealso have that property. The counterpart in a possible world ofan object in the actual world resembles the actual object, andnothing else in the possible world resembles it more. Feldman,in [11], has shown that Lewis's proposal suffers from certainserious defects. But more important, it seems to merely reraisethe key problem here. As Lewis himself notes, resemblance incertain respects is obviously going to count more than resemblancein other respects. But what makes one resemblance count more


9/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 359than another? Surely, it is going to be that certain respects aretied up with the object's very nature. But then it is going to bethe very same grounds which would determine what are the es-sential properties of the object. So we get nowhere in trying tounderstand the essential-accidental distinction by appealing tothe idea of a counterpart.All of these problems disappear if we adopt the theory ofessentialism suggested in the first part of this paper. On ourtheory, it is not necessary that an object have a property in allpossible worlds in order that it have it essentially. All that isrequired is that it have it in the actual world now and in all possiblefutures in which it exists. Or, to put it another way, a has P essen-tially just in case a has P and, for all possible future worlds in whichthere is some 01 identical with a, 01 has P. Now the only notion ofidentity involved in such essentialist claims is that of the identityof some possible future object, some object that exists in the futurein a world whose past and present is identical with the past andpresent of the actual world, with an object that exists now. Forsuch cases, we can just employ the normal criterion of identitygoverning actual objects of that type.7 Interestingly enough,Kaplan has noted that

When the worlds have a common part, as when we are consideringalternative futures to the present, the individual(s) can be tracedback to the commonpart by the usualcontinuityconditions and therecompared. ([12]: 224)He fails to notice, however, that whatever might be required tomake sense of all quantified modal claims, nothing more than thislimited notion of identity through possible future worlds is requiredto make sense of essentialist claims.

IIIWhat properties does an object have essentially? I want tocontrast three possible answers to this question: (1) it has essentiallyall those properties that it has in all possible worlds; (2) it hasessentially all those properties that it actually has and would havein all possible worlds in which it exists; (3) it has essentially allthose properties that it actually has and would have in all possiblefuture worlds in which it exists. It seems to me that (1), togetherwith certain additional assumptions, lies behind Kaplan's theoryof essentialism and its answer to our question; that (2) lies behind


10/15

360 NOUSPlantinga's theory of essentialism and its answer to our question;and that (3) lies behind the theory we are presenting here and itsanswer to our question.For Kaplan, to say that a has P essentially is to say that thereis some name a which necessarily denotes, i.e., denotes in allpossible worlds, a and which is such that rPN-lis necessarily true.As Kaplan points out, this entails that only abstract objects, whichexist necessarily, can have essential properties, for no name cannecessarily denote an object that only exists in some possibleworlds. Now this result is certainly very strange. Even if one sharesKaplan's lack of intuitions about what properties things like peoplehave essentially, it certainly seems clear that each person, like anyother object, is necessarily (and essentially) self-identical. After all,it is not merely an accident that I am self-identical.What would lead someone to such a strange theory of essen-tialism? If one believed both (1) and the often-believed8 claim thatnon-existent objects have no properties, then one would be ledto such a theory. After all, given that assumption about non-existent objects, an object that fails to exist in some world has noproperties in that world and cannot therefore, by (1), have anyproperty in any world necessarily or essentially. I suggest thereforethat these two beliefs lie behind Kaplan's theory and that whathe did was to find some de dicto reinterpretation of modal claimsthat satisfies the consequence of his assumptions, viz., that onlynecessarily existing objects have properties essentially. In order,then, to avoid his theory of essentialism, one must drop at least oneof his assumptions.It would be interesting to see what type of theory of essen-tialism would emerge if one were to drop the assumption that non-existent objects have no properties while retaining (1). It would alldepend, of course, upon the view one adopted about the propertieshad by non-existent objects. If one felt that they only had self-identity as a property, then the resulting theory of essentialism,while richer than Kaplan's theory (since it would allow for concreteobjects having some essential properties), would be weaker thanPlantinga's theory (which seems to allow for their having moreinteresting essential properties). But if, more plausibly, you allowedfor Pegasus's being a winged horse even in worlds in which he doesnot exist, then the resulting theory might well be richer thanPlantinga's theory. The trouble is that, lacking some good theory,or even some good intuitions, about what properties are had bynon-existent objects, it is difficult to see exactly what such a richer


11/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 361theory of essentialism based upon (1) would be like. It is also veryunclear as to whether one could find, for such a theory, a de dictoreinterpretation of essentialist claims.We turn to Plantinga's theory of essentialism which, as hemakes clear at the beginning of "World and Essence", is basedupon (2). But before looking at its answer as to what propertiesobjects have essentially, we must note one important difficultywith the theory as he presents it. Plantinga supposes that hisintuitions about essentialism are captured by the following de dictoreinterpretation of modal claims: an object a has a property Pessentially just in case it has P and the proposition expressed byra has non-P1, where ran is a proper name of a and rP1 a propername of P, or the proposition which would be expressed by ra hasnon-P1, if ran were a proper name of a and rFP were a propername of P, is necessarily false. It is difficult to see why he thinks so.After all, he himself says that Socrates has essentially the propertyof being a non-number, and, given (2), one certainly sees why hethinks so. But is the proposition expressed by rSocrates hasnon-P1, where 'Socrates' is a proper name of Socrates and rP1is a proper name of the property of being a non-number, necessarilyfalse ? Certainly, that sentence can be used to express a truth pro-viding that 'Socrates' is also the proper name of some number(and we can easily make it one). And unless one held the viewthat proper names have meanings, ones that vary as the referencedoes, then the proposition in question would also be true on thatoccasion.9 Plantinga, if he is to stick to (2), must therefore either(a) drop his de dicto reinterpretation of modal claims, or (b)adopt the implausible view that proper names have meanings thatvary as their reference does, or (c) switch from propositions (whichwe can think of for now as ordered pairs of sentences and meanings)to statements (ordered triplets of sentences, meanings, and assign-ments of objects and properties to the non-logical expressions inthe sentence).10Given such an account, based upon (2), what properties doobjects have essentially? Plantinga tells us that Socrates, besidestrivially having with all other objects such essential properties asbeing colored if red, has such properties as being identical withSocrates, being a non-number, and being possibly conscious essen-tially. He also argues that Socrates is essentially not a materialobject. I want to concentrate on the claims that he is essentiallya non-number and essentially possibly conscious. What supportdoes he offer for these claims and others like them ? He has no

4


12/15

362 NOUSarguments and seems to depend solely on his intuitions. Butcouldn't I imagine myself, in some possible world, as a number oras a stone which could not be conscious ? Is there anything inco-herent with that supposition ?I do not see how we can possibly answer these questions. Thereason for this is very simple: we lack an account, or even somegood intuitions, as to what trans-world identity means in a contextin which the worlds in question are non-overlapping. If we knewwhat it was for an object 01 in one possible world to be identicalwith an object 02 in another possible world, we could see whetherol's being a person and 02's being a stone, or being a number,blocks their being identical. Lacking such an account, however,we cannot answer these questions and others like them, and cannottherefore tell whether Plantinga's essentialist claims are corrector not.To be sure, Plantinga feels that there is no problem abouttrans-world identity. While I certainly think that he is right indismissing epistemological versions of that problem, I cannot seehow he can dismiss the question of what trans-world identity,through non-overlapping worlds, means. For as we have justseen, in one example that can be generalized, without an answerto that question, we are not in a position to evaluate essentialistclaims based upon (2).We turn finally to our theory which is based upon (3). Here,we do not have to depend upon either a theory or intuitions aboutthings like the properties of non-existent objects or about identitythrough non-overlapping worlds. Instead, we merely have to dependupon the very familiar and comfortable distinction about the typesof changes discussed in the first section of this paper.What properties are, on this account, essential properties?Well, there are those had by all entities, existence and self-identity.Nothing can continue to exist while losing one of these properties.In addition, you and I are essentially people, that on which thewords that I am now reading is written is essentially a piece of paper,etc. And this is just what is wanted, for the essential property thateach of these objects has corresponds, as it should, to the sort ofthing the object is, to our intuitions about the object's Aristoteliansecondary substance.This last point is extremely important. In most recent discus-sions of essentialism, the whole emphasis has been on the connectionbetween essential properties and properties that, in some sense oranother, the object must have. But there is traditionally this other


13/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 363important connection, the connection between the properties thatone has essentially and the kind of thing that one is. Traditionally,the kind of thing that one is is determined by the set of one's essentialproperties. And it is therefore an advantage of our theory ofessentialism that it have as essential the properties that are neededif our kind is to be people, etc.This leads us to an extremely important question: is it the case,for every property P that is had essentially by some object, that allobjects that have it, have it essentially, or are there some propertieswhich are had by some objects essentially and by others acci-dentally? We shall say that any property had essentially by someobject and accidentally by none, whether actual or potential,determines a natural kind, and that the set of objects having thatproperty is a natural kind. Then the question that we have raisedreally is the question as to whether, on our account, all essentialproperties (properties had essentially by something) determinenatural kinds ? It would be very good if the answer to this questionwere yes, for then the connection between essentialism and thisidea of sorts of things would be greatly strengthened.Unfortunately, our theory of essentialism, like all others, runsprey to the following example of an essential property that is not anatural kind first raised by Plantinga ([17]: 465): let P be someproperty had essentially by a but not by b, and let Q be someproperty had accidentally by b. Then the property of being either Por Q is had essentially by a and accidentally by b. So the claim thatall essential properties determine natural kinds must be false.It is important to note that this argument of Plantinga's isperfectly general and applies equally well to any theory of essen-tialism. Is there any way to meet it? As far as I can see, the only wayto do so would be to deny that a has the disjunctive property essen-tially just because he has one of the properties in question essen-tially. One way, but certainly not the only way, to do this would beto argue that, although there are such things as disjunctive predi-cates, there are no such things as disjunctive properties, and, afortiori, disjunctive properties had essentially. But of course,saying this without some general theory of propertyhood to backit up is merely expressing a pious hope (which is not necessarilyto say that it is a bad thing to do).There is another way in which the connection between essentialproperties and the sort of thing that one is breaks down. Considerthe property of being the person who was in place P1 at time t1 andconsider Joe who has that property. On our account, Joe has that


14/15

364 NOUSproperty essentially, for there is no alteration that he can undergowhich would involve his losing it; he loses it, if at all, only whenhe stops existing. And yet, one would not have supposed intuitivelythat that property has anything to do with determining the sort ofthing that Joe is.Once more, there is a way to avoid that problem. One can saythat while there certainly are such things as tensed predicatesand while there certainly are states of affairs which consist in havingcertain properties at certain times, there are no such things astensed properties, and, a fortiori, no such things as tensed propertieshad essentially. But once more, such a claim would need the backingof a theory of propertyhood.Of course, one might at this point be tempted to concludethat there is no necessary connection between essential propertiesand natural kinds or, at least, that there is no such connection inthe case of at least some essential properties. This would naturallygive rise to the question as to which essential properties determinenatural kinds and which do not.1" But I should like to avoid thisconclusion; after all, there are other reasons for being unhappywith such things as tensed and disjunctive properties. Still, whichmove should be made here is a matter for further research.I conclude, then, that it is possible to construct a straightforwardtheory of essential properties that is not open to any of the standardobjections raised against essentialism and which does allow forsome interesting essentialist claims, including, most importantly,those that provide a connection between properties had essentiallyand the sort of thing that an entity is. But if that connection isto be just as we would like it, we require some theory of property-hood that enables us to rule out some undesired essential propertieson the grounds that they are not properties at all.

REFERENCES[1] Aristotle, Physics.[2] , On Generation.[3] Brody, Baruch, "De Re and De Dicto Interpretations of Modal Logic,"Philosophia (1972).[4] ,"Is There a Philosophical Problem about the Identity of Substances ?",

Philosophia (1971).[5] , "Locke's Theory of Personal Identity," American PhilosophicalQuarterly (1972).[6] , "Natural Kinds and Real Essences," Journal of Philosophy (1967).[7] , "Two Recent Contributions to the Theory of Reference," Studies inthe History and Philosophy of Science (1971).


15/15

GOOD OLD-FASHIONED ARISTOTELIAN ESSENTIALISM 365[8] Camp, Joseph, "Plantinga on De Dicto and De Re," THIS JOURNAL, 5 (1971).[9] Cartwright, Richard, "Some Remarks on Essentialism," Journal of Philosophy

(1968).[10] Chisholm, Roderick, "Identity Through Possible Worlds: Some Questions,"THIS JOURNAL, 1 (1967).[11] Feldman, Fred, "Counterparts," Yournal of Philosophy (1971).[12] Kaplan, David, "Quantifying In," reprinted in Davidson, D. and Hintikka, J.,editors, Words and Objections(Dordrecht: Reidel, 1969).[13] Lewis, David, "Counterpart Theory and Quantified Modal Logic," Journal ofPhilosophy (1968).[14] Marcus, Ruth Barcan, "Essentialism in Modal Logic," THIS JOURNAL, 1 (1967).[15] Plantinga, Alvin, "De Re et De Dicto," THIS JOURNAL, 3 (1969).[16] , What George Could Not Have Been," THIS JOURNAL, 5 (1971).[17] , "World and Essence," Philosophical Review (1970).

[18] Quine, W. V. O., Word and Object (Cambridge, Mass.: M. I. T. Press, 1960).[19] Thomason, Richmond, "Species, Determinates, and Natural Kinds" THISJOURNAL, 3 (1969).

NOTES'In [6] and [7]. I have elaborated upon the parts of it discussed below inSection II in [3].2 Providing, of course, that one has a suitably broad understanding of eitherwhat are the properties of abstract objects or of what is a change. If one supposes,however, that changes only involve properties and that not every predicate deter-

mines a property, then this remark may need modifying.3 That of Kaplan in [12] and Plantinga in [15] and [17].4Quine [18] recognized that the first could be dealt with and thereforeproposed the second objection. Plantinga [15], Cartwright [9], and Marcus [14]have shown what is wrong with the second objection.5I have argued elsewhere, in [4] and [5], that there are problems with thisstandard claim. But even if the alternative, sketched in [4], is adopted, the samepoint can be made about it. On the possible additional requirement, see my [6].6 As is suggested by some formulations of the problem in [10] and [17].7Actually, in order to keep full transitivity of identity, we have to stipulatethat the future objects are identical just in case they are identical with the sameactual object.8 Although, I think, false. On this point, see my [7].9 This objection is similar to the one raised by Camp [8]. But he set up hisexample in such a way that Plantinga [16] was able to view the proper name inquestion as a disguised definite description. That option does not seem open to himhere.10 The de dicto reinterpretation would then be: a has P, and the statementexpressed by Fa has non-P1, where Fa' is a proper name of a and is used to refer toa and FIJP is a proper name of P used to refer to P, or the statement that would beso expressed if Fan and FIJP were to meet these conditions, is necessarily false.11If one were to adopt this approach, it would be useful to begin by consideringthe fruitful suggestions made by Thomason in [19].

Documents

Brody 1973 Why Settle for Anything Less Than Good Old-Fashioned Aristotelian Essential Ism