Within the general framework of the theory of direct reference, there is no agreement as to whether unactualised possible objects (from now on, possibilia) can be referred to by means of directly referential singular terms (from now on, DR terms). While some have maintained that such a direct reference can be established e.g. via some fixing-reference description (Kaplan, Salmon, and perhaps Kripke himself), others have denied any such possibility. In what follows, I will scrutinise such denials by attempting at the same time to provide some counterarguments to them. Indeed, I believe that possibilia can be referred to directly, primarily if the appropriate fixing-reference description is provided.

1. As far as I know, three kinds of arguments have been provided in order to deny that direct reference to possibilia may obtain. These are:

i) there is no causal link between a possible object (if there are any) and the users of the DR term intended to designate it;⁽¹⁾

ii) insofar as naming is the standard case of direct reference, it presupposes the actual existence as well as the givenness of the nominatum, in order for this to be named within a linguistic institution⁽²⁾;

iii) a possible object cannot be directly referred to via a fixing-reference description, even one expressing an individual essence, because one cannot distinguish between such an object and the property expressed by this description.⁽³⁾

To start with, i) presupposes the validity of a particular version of the theory of direct reference, namely the causal theory of reference. But this theory can be questioned already insofar as real objects are at stake - to say nothing of the entities which apparently are directly referred to but cannot be involved in any causal process because of their abstract nature.⁽⁴⁾ If we focus on the core of the causal theory of reference, namely the doctrine of the 'initial baptism', there are at least two problems for it. First, when an unnoticed switch of individuals occurs at the time of baptism, a certain object can be causally en rapport with the baptizers although the name they use ends up to be the official name of another object; thus, semantic reference to this latter object is apparently not grounded causally. Second, insofar as a past object is concerned, we may use a name to designate it although this was not the term used to refer to it when it was causally available (so that no 'chain of transmission' of the name can lead one back to the object itself).⁽⁵⁾

Suppose however that the causal theorist found an answer to these problems, e.g. by dealing with the former as a special case of a referential shift⁽⁶⁾ from baptizers' (speakers') reference to the community reference, both depending on a (albeit different) causal source, and solving the latter by appealing to an identity "a=b" as believed by the users of the new DR term "a" at issue, where "b" is an old DR term causally grounded in the no longer existing object.⁽⁷⁾ Nevertheless, it would remain that to say that a causal link with an object is sufficient for directly referring by means of a singular term to it would not yet be to say that it is a necessary condition. Indeed, nothing apparently prevents someone from providing a non-causal answer for the problem the causal theory is called

next page

  contents
 

upon, that is, not the problem of how a term refers to its referent - we assume that here the answer simply is: directly - but rather the higher problem of, in virtue of what a term directly refers to its referent⁽⁸⁾. Such a non-causal answer is precisely what Kaplan gives when he argues for a reference-fixing of a DR term by description⁽⁹⁾. Thus, either the causal theorist has a successful independent argument to the extent that such a non-causal fixing does not secure direct reference to an object, or her possible appeal to our intuitions on the matter⁽¹⁰⁾ risks being question-begging. We may therefore remark that, as it stands, i) is not sufficient to rule out the possibility of direct reference to possible objects. Let us therefore pass to evaluate ii).

2. First of all, we may notice that the term "givenness" used in ii), which of course alludes to an epistemical notion, must mean "perceptual givenness". In fact, we must find a notion of epistemic givenness which entails the (actual) existence of the object given, for such an entailment seems to be required by ii). Now, the only epistemic givenness that satisfies such a requirement is that of perception (as well as that of those states ontologically founded on perception, such as e.g. memory and knowledge).

Adding perceptual givenness to existence qua condition for direct reference is manifestly intended to fix some constraints in order for an expression to really work as a DR term. To come back to a point already noted in Section 1, according to ii) it simply cannot be the case that, insofar as they are no longer available perceptually, past objects are directly referred to by means of a certain singular term (at least in absence of any suitable 'chain of transmission' for this term). Analogously, we may suppose that even the other objection formulated in Section 1, namely that of the babies' switch at the time of baptism (which makes the unperceived baby the official bearer of a certain name), can be disposed of by the sustainer of ii). Suffice it that she is ready to re-shape it as follows:

ii') directly referring to an object, either by an official name or not, presupposes the actual existence as well as the perceptual givenness of such an object.

Sub ii') eliminates the problem of institutional naming, by restricting direct reference primarily to the actual objects perceptually given. It indeed says that a DR term cannot refer but to the individual perceptually given to those who (originally) use such a term. In order for direct reference to something to be performed, it is thus sufficient that nicknames are used, or also indexicals when proper names are missing.

But so formulated ii') might be accepted by a causal theorist as a mere constraint on her own position; only objects one is in perceptual contact with can causally ground direct reference to them. Thus, if the sustainer of ii') intends to qualify her position as something more than a restrictive variant of the causal theory of reference, two moves are available to her. Either she claims that there are cases in which directly-referred entities are perceived without their being causally given or she holds

previous page

next page

  contents
 

that being perceptually given is still irreducible to being causally given. Let us scrutinise both strategies.

As to the former strategy, concerning potential examples of perceived entities that cannot be causally given, one might try to provide three kinds of entities at least: abstract objects, fictional beings, sense-data⁽¹¹⁾. However, both free idealities, such as numbers, and fictional beings which, if they are entities at all, are abstract entities as well⁽¹²⁾, can hardly be considered as perceptible⁽¹³⁾. Besides, in a broad philosophical tradition sense-data are taken as what a subject has, not as what he perceives⁽¹⁴⁾. Thus, either this tradition is wrong (which seems to me hardly plausible) or it is not perceptual givenness that allows us to directly refer to sense-data.

Now a sustainer of ii') may accept the above remarks as indicating some further constraints for direct reference, to the effect that only perceivable spatiotemporal entities are what can be directly referred to. But then it would turn out again that the entities which can be given perceptually are only those which have a causal power on their perceivers. Thus, the only chance one has in order not to reduce being perceptually given to being causally given is to perform the other strategy previously outlined, i.e. that of supplying the former notion with some additional content. This can be done e.g. if one says that perceiving a spatiotemporal object amounts to keeping track of it within the flow of its appearances. This thesis can be formulated as the following epistemical thesis: insofar as an object is perceived, it is not exhausted by any of its appearances. For any of such appearances, there may still be another subsequent appearance under which the object may be perceived.

This epistemic thesis has also a linguistic counterpart, which appeals to the definite descriptions one gives of the object perceived rather than to the aspects by means of which it presents itself to someone. For any such description, another description is always available to us which has the object involved as its denotation, by characterising it under a epistemically new aspect. By way of explanation, if I describe the object I see as "the lime tree in my garden", as "the plant closest to the fence" etc., I can also describe it e.g. as "the tree where a cuckoo has built its nest", etc. Thus, one may say that ii') amounts to the following thesis: insofar as for any such description and a perceived object one can distinguish between the description's having a denotation and the object itself, one can directly refer to the latter. But if this is the way in which one gives perceptual givenness a content beyond causal givenness, then ii') ultimately is nothing but a species of iii), in that, as we will see below, it constitutes a way of formulating the epistemological variant of the latter. But this ultimately means that iii) is the only relevant argument we have to ponder in order to examine the validity of the thesis that possibilia cannot be directly referred to.

Before we turn to iii), however, we must note that, be that as it may with givenness, both ii) and ii') state the object's actual existence as a condition for direct reference. One might thus still argue for it independently, by simply focusing on the trivial fact that direct reference to an object is a two-termed relation between a certain word and a certain object. In actual fact, however, insofar as we have no

previous page

next page

  contents
 

motive to sustain the view that the referring-relation between a word and an object is a causal one, there is no pressure to say that both terms of this relation, particularly the latter, must exist⁽¹⁵⁾. As a matter of fact, we have a lot of relations whose members are not constrained to be actual ones. Take for example the relation of being more studied than⁽¹⁶⁾. Thus, either we have independent grounds for giving up reference to possible nonexistent objects, so that we must again turn to evaluating iii), or we must acknowledge that the reference-relation belongs to the kind of relations obtaining between existent and nonexistent members. This acknowledgement can be better motivated if we consider the relation of reference as a word-entity relation which ultimately obtains in virtue of the corresponding intentional referring-relation subsisting between a certain word's user and an object. Such a relation between a subject and an object surely presupposes that the former member of the relation exists, insofar as referring is a (part of a mental) act of hers, but not that the latter exists⁽¹⁷⁾.

3. Coming thus to thesis iii), it must be primarily noted that, as it has been formulated at the beginning of this paper, this thesis corresponds to the argument against the strong version of the theory according to which direct reference to possibilia can be assessed by means of a fixing-reference description which lacks a denotation in the actual world. But there is also a weak version of the theory, which distinguishes itself from the strong version in that for it one such description which fixes the reference for a DR term gives only a sufficient (in actual fact, the most basic sufficient) condition rather than being a description yielding both a necessary and sufficient condition for the existence of the possible object involved⁽¹⁸⁾. Both descriptions may have only one denotation, if they have any; but where the latter is a de facto rigid description, in that it denotes the same individual in every possible world in which that individual exists, the former is not such, since it only denotes the same individual in every possible world in which it denotes anything at all. The two kind of descriptions therefore distinguish themselves as follows. It holds true of the latter but not of the former that there are some possible worlds in which the description lack a denotation, although the only object which can be its denotation in further possible worlds exists⁽¹⁹⁾.

Suppose for example that two such descriptions respectively are: "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth" and "the only person resulting from the union of sperm p and egg e" (this time, a certain Philipean and a certain Elizabethan gametes). As everybody knows, Queen Elizabeth I actually has no son at all, a fortiori from Philip II either. Thus, both description are actually non-denoting terms. Nevertheless, they behave differently as far possible worlds distinct from the actual one are concerned. The former description denotes merely in every possible world in which it denotes anything at all the same possible son of the two Majesties who is responsible for the incestuous generation springing off its gamete a and another gamete b of Elizabeth. Indeed, we are asked to imagine a possible world in which that son exists, but in which that incest with Elizabeth is not committed⁽²⁰⁾. On the other hand, the latter description cannot fail to have one and the same possible individual as its denotation in every possible world in which this individual exists.

previous page

next page

  contents
 

This difference aside, however, both versions of the theory agree with their detractors in preliminarily ruling out of consideration empty descriptions possibly denoting which fail to satisfy either condition. A description of this kind may be the simpler "the son of Philip II and Elizabeth I". For since this description, like the more famous "the winged horse", may have a different denotation in each different possible world, it does not specify a criterion of identity for the possible denotation(s) it picks up. Hence, it cannot fix any direct reference by means of a singular term to it (or them?)⁽²¹⁾. Thus, no weaker version of the theory according to which one may fix descriptively direct reference to possibilia will be considered here. For we easily admit from the very beginning of this discussion that any such version, which puts a description utterly failing to provide a criterion of identity for its possible denotation into issue, is patently false. No reference to a possible object can be achieved via such a description, for there is no object which so to say waits for being picked up by that description.

Thus, let us start from the weak version of the theory, and consider the modified form the corresponding objection takes:

iii') a possible object cannot be directly referred to via a fixing-reference description, even one which only gives a sufficient (or better, the most basic sufficient) condition for the existence of its possible denotation, because one cannot distinguish between such a denotation and the property expressed by this description.

Let us see how the argument for iii') runs. With respect to a possible world W' in which the description has, unlike another possible world W, no denotation, how can we single out the object which was in W the denotation of the description, since it has lost the qualification which, as regards W, uniquely identified it for us? And if we cannot, how can we say that insofar as W was concerned such a description has fixed a referent for a DR term? Staying with our example, suppose that, unlike in W, in W' no such incest from Elizabeth I and an unfortunate son of hers obtains. Suppose that, moreover, insofar as W was concerned we decided to call "Oedipus" the incestuous son of Elizabeth and Philip there denoted by "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth". As in W' however no such fruitful incestuous intercourse obtains, with respect to it how could we pick out an individual such as Oedipus, since he would have lost there the only feature that enabled us to identify him? And if we cannot, how can we maintain that "Oedipus" worked as a DR term?⁽²²⁾

I find that, as it stands, such an objection does not work. First of all, if, in order to single the same object out insofar as another possible world distinct from W is concerned, what we need is simply another description which denotes it at that world, we may construct a world-indexed description such as "the son of Elizabeth I and Philip II who is such that a sperm a of his united itself with an egg b of Elizabeth in W". This description obviously picks up the same denotatum in all possible worlds (as is the case with, say, "the actual first Protestant Queen of England"). Moreover, as regards already any actual object whose name has been fixed descriptively, we can well imagine that, for all truths actually concerning this object, each of them might have been false, without our having to

previous page

next page

  contents
 

suppose that we would have to re-identify such an object with respect to the metaphysically possible world called into play by this imagination.

But this indicates that the objection we are here concerned with fails in a substantial way. In general, if as regards W the reference-fixing of a DR term has been successful, we have indeed no reason to doubt that such a term does not refer to the same object as regards W'. If as regards a possible world W such a fixing has indeed been successful, then as regards such world we have been able to pick up an individual there existing and therefore to directly refer to it by means of such a term. But if this is the case, then we have no need of identifying that individual again with respect to another possible world W', insofar as any such world is a world whose domain has been stipulated. Here indeed if the fixing has been successful then actual and possible objects must be on a par⁽²³⁾. The only difference between the case involving realia and that involving possibilia, insofar as fixing reference is concerned, is that between singling an individual out in a world and singling it out with respect to a world. This difference, I would say, merely rests on the existential situation which concerns us qua speakers of our actual language. I allude to the trivial fact that our language, with all its referential features, is spoken by us in the actual world, the world in which we actually live. Thus, as inhabitants of this world who are as actual as the objects contained in it, we single such objects out in it insofar as we look in it for denotations here of certain descriptions belonging to our language. Concerning worlds which we do not inhabit, we single out objects existing there insofar as we can only look with respect to these worlds for denotations there of descriptions belonging to the same language. But such a difference is totally irrelevant for our concerns. We only need to select a certain world, be it the actual world or a possible one, as a world such that the relevant kind of descriptional singling has successfully fixed the reference for a DR term. Once such a world has been provided, then as to the nearest possible world one does not have to single out again the referent for such a term, insofar as it is stipulated that that world has such an object in its domain. Once we single out in the actual world Elizabeth I and refer to her by "Elizabeth", we might wonder how things might have been if Elizabeth had married Philip II. Precisely in the same sense, once we have singled out as to a world W Oedipus and we refer to him by "Oedipus", we might further wonder - an iterated counterfactual hypothesis, so to say - how things might have been if Oedipus had not committed a certain fertile incest with Elizabeth, etc.⁽²⁴⁾

However, as regards the weak version of the theory of direct reference to possibilia, we must acknowledge that there is a problematic point. But this concerns ontology rather than epistemology, which is what we have hitherto been concerned with. Recall that we have at our disposal a description, "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth", which gives by itself a mere sufficient condition⁽²⁵⁾ for the existence of its denotatum. Oedipus, we are told, is the individual which that description picks up at a possible world W. Moreover, there is a further possible world W' in which Oedipus exists without satisfying that description. But suppose that in looking among worlds for the unique denotation of that description we had started from W' rather than from W. Ex hypothesi, we would not have found in

previous page

next page

  contents
 

W' any individual satisfying it. How could we have assessed, then, that an individual such as Oedipus belonged to W's' domain? Who is Oedipus, in W'? And if we had failed to assess this, once we had passed to consider W and we had found that in W such a description has a denotation, how could we have assessed that that denotation is the same Oedipus as that existent in W'?

The problem is then (again) a problem concerning the identity of possible objects. To put it in a different way, the fact is that in order for an object (indeed, an actual as much as a possible one) to be the referent of a DR term with respect to a certain possible world, one must have it at one's disposal as a genuine member of such a world before it is assessed as the satisfier in that world of a description which gives a sufficient condition only for its denotatum's existence. And to have an object at one's disposal presupposes that it satisfies a (trans-worldly) criterion of identity, so that no problem such as those of the previous paragraph remain unanswered.

Let us thus say the following. For a description such as "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with a certain egg b of Elizabeth" it is possible that it has a (only) denotation; but for such a description there must by itself be no object which possibly is its (only) denotation. In more formal terms, once "d" replaces "description" and "D" "denote", where:

(1) $d ŕ ($!x (E!x & dDx))

is true,

(2) $d r ($!x ŕ (E!x & dDx))

is false (recall that in (1) and (2) the existential quantifier is to be meant non-existentially, in that it quantifies over possible objects, and "E!" means existence taken as a first-order property.⁽²⁶⁾

The point is thus that, seen this way, the problem raised by a description such as "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth" is not essentially different from that, already seen, raised by actually unsatisfied descriptions that may be satisfied by more than one thing. As we said, insofar as a description such as "the winged horse" or even "the son of Philip II and Elizabeth I" is plurally satisfiable, it does not provide a suitable criterion of identity; thus, its having a possible denotation cannot be counted as its denoting a particular possible object. Now, if we attach to such a description the qualification "who is such that a certain sperm a of his united with a certain egg b of Elizabeth" the situation does not change so much. Sure, unlike "the son of Philip II and Elizabeth I", it is not the case that such a description may be satisfied by more than one thing. Nevertheless, insofar as there are possible worlds in which it is not satisfied even though what allegedly is its only denotation in other worlds exists, it does not fix a given possible individual as a direct reference for a name to which it is correlated⁽²⁷⁾.

It may well be that the phrase "sufficient conditions for the existence of its denotation" here leads one astray, so that one may still consider e.g. "the son of Elizabeth I and Philip II" and "the son of

previous page

next page

  contents
 

Elizabeth I and Philip II who is such that a certain sperm a of whom his with a certain egg b of Elizabeth" as belonging to irreducibly different kinds of description. Nevertheless, suppose that (to simplify matters) one passes to the example provided by M. Cook⁽²⁸⁾ as a case of a description which designates the same individual wherever it denotes anything at all, namely "the father of Elizabeth II", which has George VI as its actual denotation⁽²⁹⁾. Now, it is certainly true that in order for someone to be George VI it is sufficient to be Elizabeth's father, in that nobody but George can satisfy the above description. But it is still not true that in order for an individual such as George VI to be, i.e. to be counted within the overall domain of what there is, it is sufficient that such a description has a unique denotation in all worlds in which it denotes anything at all. Of such a description, in fact, it is true that it may have only one denotation, but it is not true that there must be an individual which possibly is its only denotation.

4. Thus, I admit that the weak version of the theory of direct reference to possibilia appears to be ultimately ungrounded for ontological rather than epistemical reasons. What about the strong version, however? Well, one must primarily acknowledge that a description which, in Monte Cook's terms, gives both the necessary and the sufficient condition for the existence of an object, does such a job of providing by itself an individual. In other words, it is a description of which (1) holds true. In fact, not only such a description may be satisfied by one object only, as was the case with the previous kind of description too, but also it is a de facto rigid description, in that it is satisfied by that object wherever the latter exists. Thus, the ontological problem envisaged in the previous Section disappears. Only one individual may satisfy that description across worlds - that is, in all possible worlds in which such individual exists - so that it cannot be the case that he is not one and the same individual across such worlds. Let us assume, therefore, that an example of such a de facto rigid description is precisely "the only person resulting from the union of sperm p and egg e (a Philipean and an Elizabethan gametes)"⁽³⁰⁾. As has been noted, one may say that such a description expresses an individual essence, namely an essence that may be possessed by one object only⁽³¹⁾. Once it has thus fulfilled the task of providing a certain possible individual, we may use this description in order to single out that individual with respect to the possible worlds in which this exists and hence to fix the reference of a DR term accordingly⁽³²⁾. However, a thesis stronger than iii'), namely iii), has precisely been put forward as a thesis against such a fixing-reference to a possible object achieved by means of such a description. I quote it extensively from its original sustainer:

"The description supposedly singles out a possible object. If it does, then we should be able to cast off the description and just talk about the object. [...] [But] with actual objects, we can get some hold of the difference between an individual's existing and a description's be satisfied. ... It is natural to see this difference as being between a more general possibility that a childless individual result from a certain sperm and egg and a particular possibility that that individual, George VI, be childless. The

previous page

next page

  contents
 

difference is between a possibility concerning a description's being satisfied and a possibility concerning an individual singled out by either the description or the name. With merely possible objects, however, such an understanding of the difference between the possibilities breaks down. We cannot contrast the more general possibility that a childless individual result from sperm s and egg e with the particular possibility that that individual, Gertrude, be childless, because there is no Gertrude and hence no separate content in the reference to that individual. There is no difference between a possibility concerning a description's being satisfied and a possibility concerning an individual singled out by either the description or the name, because there is no individual singled out and detachable from the description." (Cook (1983:308))

Now, there are again two ways in which such an argument can be meant: one epistemic and the other ontological. Let us start with replying to the epistemic reading. Once we have at our disposal an individual - and now we have one, insofar as we have in possible worlds different from the actual one the instantiation of a certain individual essence - we can refer to it by means of a DR term fixed by such a description. To be sure, since it does not actually exist we cannot perceive it. Thus, we cannot discover any new aspect of it, so that we apparently have to stay with its original, essential, characterisation. Still, we may have at our disposal a host of world-indexed descriptions, such as "the F in W", as well as a lot of intentional descriptions (that is, descriptions which express converso-intentional properties, such as "the thought-of F"), which help us to characterise the same possible entity in a different way, insofar as we can no longer doubt whether they pick up the same denotatum or not once this is individuated by its individual essence. But let us put this aside. My question rather is: why should our being epistemically linked to an essential characterisation of a possible object be relevant insofar as referring to it is concerned? That we are not able to discard its essential characterisation has nothing to do with referring to it, at least insofar as from an ontological point of view there is a difference between it - an individual, albeit possible - and its essence - an individual concept.

Now, it is easy to see that such an epistemic version of iii) is precisely the interpretion of ii') we left open in Section 2 for further discussion. Thus, in order to defend the epistemic version of iii), one might perhaps exploit what it turned out in that Section as regards perceptual givenness by taking it as a sort of transcendental argument against direct reference to possibilia. That is, one might assert that directly referring to an object is possible only insofar as one may keep track of the referent within the flow its modifications. As we saw, this is what gives perceptual givenness its specific content⁽³³⁾.

In order for this argument to be a transcendental one, it should claim that its alternative - namely, that one can directly refer to a possibile without discarding its essential characterisation - is unconceivable. This claim would makes the argument undoubtedly stronger for instance than thesis i), whose main flaw, as we saw, was precisely that it did not say anything as to why non-causal direct reference was to be ruled out. Still, I cannot see any reason why such an alternative should be unconceivable. On the contrary, it seems to me quite understandable. Again indeed, one thing is to characterise an object by means of its essential features, another one is to refer to it by means of a

previous page

next page

  contents
 

DR term. Such a difference does not depend on whether the object actually exists or not. If the object does not exist, then it is true that we cannot discard its essential characterisation from it. However, this does not make referring to it by means of a DR term collapse on to essentially characterising it, as the transcendentalist claims. Such an epistemic impossibility of discarding rests on the mere contingency that we live here, in the actual world, not in the world(s) in which such an object exists. Had we lived there, we would have been able to provide that discarding in the same way as we actually do with actual objects. And this ultimately implies that we would have been able to single that object out in that world, as we do with actual objects: normally, by discriminating them perceptually. Now, insofar as we, as a matter of fact, do not inhabit that world, but the actual one, once the existence-entailing character of perception is also assumed it trivially follows that we cannot discriminate perceptually objects which do not exist in the actual world. Thus, as I said before, we can only single possible objects out with respect to, or at, those possible worlds in which they exist.

At this point, either our transcendentalist shows that the notion of singling objects out at a world betrays some conceptual incoherence, or again his position is hardly maintainable. I cannot see, however, any incoherence in such a notion. I do single out different possible objects even at the same possible world, that is, even when these objects are compossible ones. Indeed, I individuate distinct possible objects by means of their different individual essences. For instance, I individuate Elip, our unactual offspring of the afore-mentioned Elizabethan and Philipean gametes, as distinct from Napoleon jr., the unactual offspring of certain Napoleonian and Josephine Beauharnais' gametes.

I imagine that a transcendentalist would reply that for her, by invoking the notion of singling objects out at a world, I have not proceded any step further, since to distinguish possible (and even compossible) objects by means of their different individual essences is still indiscernible from distinguishing these individual essences themselves. But suppose we take a plurality of possible worlds into consideration. We may say that, once we have a world W in which an individual essence, actually unsatisfied, is instantiated, we may have a lot of other metaphysically possible worlds W', W"... in which what uniquely instantiatiates that essence in W as well as in all such worlds is subject to all conceivable variations. For instance, we may suppose that the unactual offspring of those Elisabethan and Philipean gametes might have reigned over Spain, or over England, or even over of both⁽³⁴⁾. But this would amount to saying that we are able to attribute to what uniquely instantiatiates a certain individual essence something more than its essential characterisation. Now, this seems be the same liability as that of attributing to an actual object something more than its essential characterisation. (We may moreover assume, as even some of the defenders of iii) acknowledge, that actual objects have an individual essence⁽³⁵⁾.) The fact is that by taking a sequence of possible worlds into consideration we are able to do with possible objects what we do with actual objects by taking a sequence of actual temporal instants into issue. By modally imagining possible variations in which a possible object is involved, we keep track of it among the flow of such variations in the

previous page

next page

  contents
 

same epistemic way in which we keep track of an actual object among the flow of its actual variations by temporally going on perceving it.

At this point, we are faced with two options. Either this analogy between possible and actual objects shows that perceptual keeping track of an object is unnecessary in order to directly refer to it⁽³⁶⁾ or it shows that, insofar as what we do while perceiving actual objects is to attribute them something more than their essential characterisation, we do not directly refer to them either. Neither alternative is obviously agreeable for our transcendentalist. If she chooses the former, then her argument fails immediately. If she opts for the latter, then she has to give up her claim that we do refer directly to actual perceived objects, so that her arguments fails again. Thus, it finally seems to me that the burden of the proof is on those who claim that direct reference to possibilia is unconceivable to show that it is really so. In lack of such a proof, the 'transcendental argument' appears ultimately to miss its target.

Thus, the last way out for maintaining iii) is to defend precisely its ontological version, namely that according to which a possible unactualised object is nothing but its individual essence. First of all, it must by now be clear that iii) taken in such a reading is quite different from the typical Quinean objection that has been raised against possibilia, viz. that according to which a possibile does not satisfy a proper criterion for its identity⁽³⁷⁾. Insofar as it is the only object which may possess a certain individual essence, a possibile does indeed satisfy such a criterion. Rather, what iii) in such a reading ultimately says recalls the Plantingian (actualist) thesis according to which insofar as possibilia are concerned we cannot have at our disposal more than an unactualised individual essence, in that to say that in a possible world an unactual object X exists, where "X" is a proper name, is to say that in such a world a certain individual essence, actually unsatisfied, is instantiated⁽³⁸⁾.

Unfortunately, such a thesis is wrong, at least insofar as even actual objects, as I have just said above, have an individual essence. In fact, let us suppose that the individual essence of, say, the present Prince of Wales, Charles, is that of being the offspring of two gametes, one Windsor and one Mountbatten. Nobody questions that Charles might not have existed, for instance in a possible world W. Suppose now that among the inhabitants of such a possible world there are some philosophers who maintain that possible objects are nothing but their individual essences. In particular, they maintain that a certain possible, there nonexistent, object which is the, there unactualised, offspring of the afore-mentioned gametes (suppose for simplicity that Elizabeth II and Philip Mountbatten still exist in W), is nothing but such an individual essence. Thus, no matter which name they had used in order to denote it, they would have employed nothing but an abbreviated description. Well, the fact that what people would do and say in such a world is, as Kripke has repeatedly said⁽³⁹⁾, of no interest to us insofar as our counterfactualisations are concerned does not prevent us from saying that, if the philosophers of W maintained the above thesis as regards the (possible) offspring of a certain Mountbatten sperm and a certain Windsor egg, they would be wrong. Nobody questions that an actual object is more than its individual essence, in the sense that the former is irreducible to the latter⁽⁴⁰⁾. Should such an individual not exist in a certain possible world, this would not mean that he has become identical with his individual essence, regardless of what people, who are able to grasp

previous page

next page

  contents
 

such an essence precisely as we do, would say in such a world. Thus, by parity of reasoning, such people could say that we are wrong if we said that a possible object is identical with its individual essence insofar as an object existing in that world although not in ours were indeed concerned. Therefore, we can ultimately say that no object, regardless of its actually existing or not, is other than its individual essence, so that thesis iii) turns out to be false even in its ontological reading⁽⁴¹⁾.

If iii) turns thus out to be false, we can legitimately conclude that objections i)-iii) can be rejected and thesis:

iv) possibilia can be directly referred to, at least by means of descriptions which express individual essences⁽⁴²⁾ and therefore provide unactual individuals to be designated⁽⁴³⁾ may be positively held instead⁽⁴⁴⁾. Moreover, sentences containing DR terms designating possibilia may be regarded as expressing singular, not general, propositions, with possible objects among their constituents. To this latter point, one might retort that there cannot be singular propositions whose members do not actually exist⁽⁴⁵⁾. Such a reply, however, does not work. A proposition is an abstract entity, and, appearances notwithstanding, for an abstract entity to actually subsist it does not need its members to actually exist in the same world. One can see this with sets: we have actually subsisting sets whose members do not actually exist⁽⁴⁶⁾. More generally, we may assert that for an abstract entity which is constituted out of other entities to subsist in a world is independent on its members' existence in that world; a member's failing to exist in a world does not strangely compel the abstract entity to which it belongs not to subsist in that world⁽⁴⁷⁾.

Before concluding, let us give a quick look at the ontological consequences of the theses defended in the last paragraph. The latter remarks can help us to reduce the gap which, as regards possible worlds, appears to obtain from a metaphysical standpoint between our position and that of actualism, i.e. the doctrine according to which there are only actual objects⁽⁴⁸⁾. On the one hand, although we are friends of possibilia, our semantical choice of sustaining the theory of direct reference plus the adoption of Kripkean possible worlds semantics⁽⁴⁹⁾ has indeed distantiated us from the Lewisian, modally realist, form of possibilism. As is well known, according to such a position possible worlds are genuine entities each endowed with its own array of individuals. This entails that any sentence containing DR terms must actually be multiply evaluated, that is, as concerning a different item for any different world.⁽⁵⁰⁾ On the other hand, the above friendship literally prevents us from making company with actualism either. Now, actualism commits one to the thesis that if there are possible worlds in which there are obtaining states of affairs (true propositions) which do not concern actual individuals, these states can only be existentially generalised states (propositions).⁽⁵¹⁾ This thesis has the unpleasant consequence that possible worlds which differ only by containing distinct possible individuals could not be discriminated, insofar as they are one and the same world-type.⁽⁵²⁾ But one may now note that this thesis is forced by the dubious assumption according to which there cannot be singular propositions whose members do not actually exist. Once we are able to remove such an

previous page

next page

  contents
 

assumption, we may tentatively share with actualists the idea that possible worlds are nothing but maximal consistent sets of propositions, insofar as we take such sets as containing also singular propositions whose members are either actual or possible existents⁽⁵³⁾. The metaphysical outcome of such a move is the defense of a Possibilist Combinatorialism, according to which possible worlds are ultimately nothing but the possible combinations of a fixed domain of possible objects, some of which actually exist and some of them do not, but might have.^(54),(55)

REFERENCES

Adams, R.M. 1977 Critical Study. The Nature of Necessity (A. Plantinga), Noûs 11, pp. 175-191.

Adams, R.M. 1981 Actualism and Thisness, Synthese 49, pp. 3-41.

Almog, J. 1991 The Subject-Predicate Class I, Noûs 25, pp. 591-619.

Baldwin, T. 1982 Sets Whose Members Might Not Exist, Analysis 42, pp. 133-138.

Barcan Marcus, R. 1985/6 Possibilia and Possible Worlds, Grazer Philosoph. Studien 25/26, pp.108-133.

Bradley, R. 1989 Possibility and Combinatorialism: Wittgenstein versus Armstrong, Canadian Journal of Philosophy 19, pp. 15-41.

Castaeda, H-N. 1985/6 Objects, Existence, and Reference. A Prolegomenon to Guise Theory, Grazer Philosophische Studien 25/26, pp. 3-59.

Cocchiarella, N.B. 1982 Meinong Reconstructed Versus Early Russell Reconstructed, Journal of Philosophical Logic 11, pp. 183-214.

Cook, M. 1985 Names and Possible Objects, The Philosophical Quarterly 35, pp. 303-310.

Corazza, E. 1989 Actualisme et possibilisme: l'évaluation dans les mondes possibles, Logique & Analyse 125/126, pp. 81-111.

Donnellan, K. 1979 The Contingent A Priori and Rigid Designations, in P.A. French et al. (eds.), Contemporary Perspectives in the Philosophy of Language (Minneapolis:University of Minnesota Press), pp. 45-60.

Evans, G., 1973 The Causal Theory of Names, Proceedings of the Aristotelian Society suppl.vol. 47, pp.187-208.

Forbes, G. 1980 Origin and Identity, Philosophical Studies 37, pp. 353-362.

Kaplan, D. 1973 Bob and Carol and Ted and Alice, in K.J.I. Hintikka et al. (eds.), Approaches to Natural Language (Dordrecht:Reidel), pp. 490-518.

Kaplan, D. 1989a Demonstratives, in J. Almog et al. (eds.), Themes from Kaplan (Oxford:Oxford University Press), pp. 481-563.

Kaplan, D. 1989b Afterthoughts, in Themes from Kaplan cit., pp. 565-614.

King, P.J. 1993 Lycan on Lewis and Meinong, Proceed. of the Aristot. Society 93, pp.193-201.

Kripke, S. 1971 Semantical Considerations on Modal Logic, in L. Linsky (ed.), Reference and Modality (Oxford:Oxford University Press), pp. 63-72.

Kripke, S. 1973 The John Locke Lectures 1973, unpublished typescript.

Kripke, S. 1980 Naming and Necessity (Oxford:Blackwell).

Künne, W. 1982 Criteria of Abstractness, in B. Smith (ed.), Parts and Moments (Munich:Philosophia Verlag), pp. 401-437.

Kvart, I. 1993 Mediated Reference and Proper Names, Mind 102, 611-628.

previous page

next page

  contents
 

Lewis, D. 1979 Possible Worlds, in M.J. Loux (ed)., The Possible and the Actual (Ithaca:Cornell University Press), pp. 182-189.

Linsky, B., Zalta, E.N. 1994 In Defense of the Simplest Quantified Modal Logic, Philosophical Perspectives 8, pp. 1-35.

Lycan, W.G. 1993 Armstrong's New Combinatorialist Theory of Modality in Ontology, Causality and the Mind, J. Bacon et al. (eds.) (Cambridge:Cambridge University Press), 3-17.

McMichael, A. 1983 A Problem for Actualism About Possible Worlds, The Philosophical Review 92, pp. 49-66.

Napoli, E. 1992 Riferimento diretto in M. Santambrogio (ed.), Introduzione alla filosofia analitica del linguaggio (Roma-Bari:Laterza), pp. 385-429.

Napoli, E. (forthcoming) (Direct) Reference.

Noonan, H. 1983 The Necessity of Origin, Mind 92, pp. 1-20.

Parsons, T. 1979 The Methodology of Nonexistence,The Journal of Philosophy 76, pp. 649-662.

Parsons, T. 1982 Fregean Theories of Fictional Objects, Topoi 1, pp. 81-87.

Plantinga, A. 1974 The Nature of Necessity (Oxford:Clarendon Press).

Plantinga, A. 1976 Actualism and Possible Worlds, in The Possible and the Actual cit., pp. 253-273.

Plantinga, A. 1979 De Essentia, Grazer Philosophische Studien 7/8, pp. 101-121.

Putnam, H. 1975 The Meaning of Meaning, in Mind, Language and Reality (Cambridge:Cambridge U.P.), pp. 215-271.

Quine, W.V.O. 1953 On What There Is, in From a Logical Point of View (New York:Harper & Row), pp. 1-19.

Recanati, F. 1993 Direct Reference (Oxford:Blackwell).

Rosenkrantz, G. 1984 Nonexistent Possibles and Their Individuation, Grazer Philosophische Studien 22, pp. 127-147.

Rosenkrantz, G. 1985/6 On Objects Totally Out Of This World, Grazer Philosophische Studien 25/26, pp. 197-208.

Russell, B. 1937 The Principles of Mathematics, Cambridge U.P., Cambridge

Salmon, N.U. 1979 How Not to Derive Essentialism From the Theory of Reference, The Journal of Philosophy 76, pp. 705-725.

Salmon, N.U. 1982 Reference and Essence (Oxford:Blackwell).

Salmon, N.U. 1987 Existence, Philosophical Perspectives 1, pp. 49-108.

Searle, J.R. 1983 Intentionality (Cambridge:Cambridge University Press).

Teichmann, R. 1991 Future Individuals, The Philosophical Quarterly 41, pp. 194-211.

Voltolini, A. 1991 Objects As Intentional and As Real, Grazer Philosophische Studien 41, 1-32.

Voltolini, A. 1994a Can Negative Existentials Be Referentially Vindicated?, Lingua e stile 29, 397-419.

Voltolini, A. 1994b Ficta versus Possibilia, Grazer Philosophische Studien 49 (forthcoming).

Williamson, T. 1990 Necessary Identity and Necessary Existence, in R. Haller, J. Brandl eds., Wittgenstein: Towards a Re-Evaluation I (Vienna:H]er-Pichler-Tempsky), pp. 168-175.

Wittgenstein, L. 1953 Philosophical Investigations (Oxford:Blackwell).

Zalta, E.N. 1983 Abstract Objects (Dordrecht:Reidel).

previous page

next page

  contents
 

NOTES

1. Cf.Teichmann (1991:204)

2. Cf. Barcan Marcus (1985/6:125-6).

3. Cf. Cook (1983).

4. For the thesis that abstract entities can be directly referred to even though they do not affect us causally, cf. e.g. Kaplan (1989b:607fn.101).

5. This case is implicitly introduced by Almog (1991:604-6). See also this remark by Kaplan: "if we were to discover that Aristotle ... had been dubbed 'Aristotle' only in medieval times, the name ... would still be a name, with all its attendant powers" (1989b:605).

6. On which cf. Evans (1973).

7. Traces of both answers can be found in Kvart (1993:611,613).

8. Cf. Almog (1991:597-9). On the contrary, treating a semantic relationship tout court in terms of a causal one may be charged of being a conceptual confusion: cf. Kaplan (1989b:605 and fn.92).

9. Cf. Kaplan (1989b:605).

10. See again Kvart (1993:613).

11. Insofar as natural kinds can be taken as instances of abstract objects, e.g. species (for such an opinion, cf. Künne (1982)), Putnam (1975) and Kripke (1980) can be held as sustainers of the extension of the theory of direct reference to abstract objects. As to the parallel extension concerning ficta, cf. Napoli (1992:389). Direct reference to sense-data, moreover, is hypothesised in Kripke (1973).

12. For such a thesis, cf. Cocchiarella (1982:213); Voltolini (1994b)

13. I have deliberately said "free idealities" rather than "bound idealities" (namely, astract objects which, unlike the former, have instances which are spatio-temporal beings; for this distinction between these kinds of abstract objects, cf. Künne (1982:408)), since one might object that bound idealities at least can be perceived insofar as they can be indicated via the demonstration of one of their instances. But the evidence for this is rather weak. To be sure, I can definitely say: "this [species] is dangerous" while pointing at some instances of the indicated species, e.g. some panthers. But one might hold that the demonstrative pointing in such an utterance is a case of deferred ostension, such that no species would be directly perceived when pronouncing that utterance. As Tim Williamson pointed out to me, one may also refer to free idealities by deferred ostension without obviously perceiving them. I might for instance say "I ate that number of chocolates" while pointing in the direction of ten wrapping papers.

14. See for instance Searle:"When I see a car, or anything else for that matter, I have a certain sort of visual experience. In the visual experience of the car I don't see the visual experience, I see the car; but in seeing the car I have a visual experience ... though the visual perception always has as a component a visual experience, it is not the visual experience that is seen, in any literal sense of 'see'." (1983:37-8).

15. On reference as being a relation which is liable to subsist actually between an existent and a nonexistent object, see also Salmon (1982:39fn.41) (1987:94). Salmon however wishes to retain this thesis while not committing himself to possible entities, by holding the (admittedly mysterious) claim that "while there is no one actually denoted by 'Noman' [i.e., the possible individual who would have resulted if sperm s had fertilised egg e], there still might have existed someone who not only would have been, but actually is, denoted by 'Noman'". Such a thesis commits him to reject what has been qualified "serious actualism" (cf. Linsky-Zalta (1994:11,15-6)). Indeed, although for him actual

previous page

next page

  contents
 

objects are the only objects there are, (nominal) properties can be truly predicated of unactual objects. Now, he motivates his thesis by saying that the committment to possibilia engages one to hold the apparently contradictory claim that there are objects that do not exist (cf. (1982:39fb.41)). As Meinong and early Russell well knew, however, there is no real contradiction in such a claim, once one distinguishes between the concept of being (actually true of both possible and actual entities) and that of existence (belonging actually only to actual entities). Cf. on this Williamson (1990).

16. For other such examples, cf. e.g. Castaeda (1985/6:13-4). See also Parsons (1979:660). Now, one might reply that insofar as a sentence such as "Mercury is more studied than Vulcan" seems to concern a non-existent entity named Vulcan, it has to be paraphrased in such a manner that the apparent reference to such an entity is eliminated. However, as Kripke (1973) has convincingly shown, any such paraphrase (notably, the "descriptivist" and the "metalinguistic" ones) do not seem to work once that sentence is evaluated with respect to possible worlds. That is to say, the sentence and its alleged paraphrase have always different truth-values with respect to the same possible world. On this, cf. Voltolini (1994a:398-401). For another critical view on paraphrasing-strategies, cf. again Parsons, who shows that sentences such as the above do not introduce a nonextensional context (1982:83-6).

17. One might still retort that, insofar as such a relation is an intentional relation, one must assume that its right member must be somehow given to the left member. I do not want to question it. I would only like to remark that, if this (as I believe) is the case, then the givenness involved must not be a perceptual one.

18. For this distinction among descriptions and the following remarks, cf. Cook (1985:305-6).

19. The differentia specifica between these descriptions is however that given in terms of the description's yielding either a sufficient or a both necessary and sufficient criterion for the identity of its denotatum. One might indeed provide "the only explorer resulting from the union of sperm s and egg e" as an example of a description which may have one denotation only whenever it has a denotation. However, such a description would also yield both a necessary and a sufficient condition for the existence of its only denotation, since it is construed out of a description which expresses those conditions, namely "the only individual resulting from the union of sperm s and egg e".

20. To be sure, in order for the description "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth" to work as a description which denotes the same individual wherever it denotes anything at all, the name "a" occuring in it must be taken as a term which rigidly designates a certain possible gamete. Here the problem is not that we presuppose what we must demonstrate, namely that names may directly refer to possible unactual objects. For argument's sake, one might even accept that "a" here merely behaves as a de facto rigid description expressing an individual essence, namely a property which gives both a necessary and a sufficient condition for being the description's (only) denotatum. The problem is rather whether we are not compelled to say that "a" replaces an utterly non-rigid (actually non-denoting) description. Were it such, the description "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth" would be such that it had different possible denotations for each distinct possible world: an incestuous son of the two Majesties one of whose sperms is responsible for the relevant incest with Elizabeth in a world W, another son of the two Majesties one of whose sperms is responsible for the relevant incest with Elizabeth in a world W', etc. Thus, it would no longer be a description which denotes the same possible individual wherever it denotes anything at all. This may explain why Cook, after having given an example of the above kind of description which actually denotes ("the father of Elizabeth II"), refuses to give an example similar to the one I have provided of an analogue description which possibly denotes. On the contrary, he utilizes the same description "the only person resulting from sperm s and egg e" twice, by supposing, first, that such a description gives

previous page

next page

  contents
 

sufficient conditions, and second, that it gives both necessary and sufficient conditions for the existence of the relative object (1983:306). However, the description I choose, "the son of Philip II and Elizabeth I who is such that a sperm a of his united itself with an egg b of Elizabeth", is such that it cannot fail to have a denotation because the unicity condition is not satisfied insofar as a certain world is made issue of. Cook's example taken in its first supposition, on the contrary, admits such a possibility: cf. (1983:306fn.10). But this supposition involves some drawbacks: on this, cf. fn. 27.

21. Cf. Kaplan (1973:505-6), (1989b:609). See also Cook (1983:304). Incidentally, this position allows us to skip evaluating the traditional objections (originally coming from Donnellan (1979)) against the idea of fixing direct reference for a term by means of descriptions such as "the perturbator of Uranus" or "the first child born in the 21^st century". Since these descriptions fail to satisfy the required conditions concerning the identity of their denotations, they fall immediately out of our present concern.

22. Cf. Cook (1983:306).

23. Pace Cook (1983:307).

24. On the question of iterated modalities with possibilia, see fn. 34 below. In a nutshell, by proposing to extend direct reference to possibilia we are trying to defend something very close to early Kripkean (1971) non-realist semantics, so as not to endorse a Lewisian modal realism as to possible worlds. This attempt is described by Plantinga (1974:128-31) as the application of Kripke's pure semantics. (As Plantinga himself (1974:129fn.1) stresses, Kripke has never repudiated such an attempt; cf. Kripke (1980:158).)

25. With the proviso stated in fn.20.

26. For the defense both of the 'being'-reading of the existential quantifier and of the first-order nature of the property of existence, cf. e.g. Zalta (1983). Now, a strong possibilist like Zalta might say that by means of Barcan Formula (BF) "ŕ $f x É $x ŕ f" plus Necessitation one may pass from the truth of (1) to the truth of (2). Note moreover that in order to endorse (BF) one needs not be a strong Possibilist: see Williamson (1990). But to assume (BF) unrestrictedly is to put the cart before the horse. From (BF) it trivially follows that if it is possible that there is a denotation for a certain description then there is an individual which possibly is such a denotation. If what I have said in the text is right, however, this individual would be an entity for which no (cross-world) identity; and this seems to me ungrounded.

27. The situation is even worse with the original example by Cook, "the only person resulting from sperm s and egg e" taken as providing sufficient conditions only for the existence of its denotatum. So conceived, such an actually unsatisfied description can fail to have a denotation because there might be two individuals which result from that fertilised egg. But, since we do not yet have them at our disposal qua (distinct) individuals, how can we establish, with respect to a further possible world in which the description is uniquely satisfied, which of them is its denotatum there? (Forbes (1980:354) would probably hold that none of them could be such a denotatum, in that he states that the property of being an identical twin is not accidental. For him, it cannot be the case that one and the same individual has that property in a world W and has not it in a different world W', so that such an individual can be the denotatum of the above description (in the intended reading) in W' but not in W. Cf. on this also Noonan (1983:18).)

28. Cf. fn.20.

29. This example is from Cook (1983:305).

previous page

next page

  contents
 

30. Cf. Kaplan (1973:506fn.19), (1989b:607-10), Salmon (1987:94). I do not want to question, here, whether there are best candidates among descriptions which provide both necessary and sufficient conditions for the existence of their denotatum, as this is not within my present purpose. If someone holds that a different kind of description better fills the need, he can well substitute his own example to that provided here. (I only say that the factual existence of monozygothical twins would legitimate such an attempt.)

31. For a criterion concerning the descriptive expression of an individual essence which individuates a possibile, cf. Rosenkrantz (1984:130).

32. This is what Salmon explicitly maintains: to suppose that we have singled out one particular possible person from all the rest simply by noting that it is the (possible) offspring of certain gametes is to presuppose that there is only one possible individual that could have sprung from such gametes (cf. (1979:708fn.5).)

33. A similar argument has been defended in Napoli (forthcoming).

34. Incidentally, it must be recalled that the counterfactual variants hypothesised in the text works under the paradigm of an iterated modality, that is, under the pattern of sentences like "It is possible that there be a person X who does not exist in the actual world, and who performs some action Y, but who might not have performed Y". Now, McMichael (1983:54) writes that such sentences pose a serious problem for actualism about possible worlds, namely for the theory according to which there is only what there actually is. He acknowledges that such a case might be adequately treated by Kripkean non-realist semantics which admitted possibilia, if only there were such semantics. He however denies this, because in his opinion possible objects do not satisfy identity criteria (ib.:62). Insofar as there are individual unactualised essences which are not Kaplanian-Plantingian haecceitates (whose subsistence he brillantly criticises; cf. (ib.:55-61)), however, such criteria are eo ipso provided. Correspondingly, Kripkean non-realist semantics may well obtain and treat with the above problem.

35. Cf. Cook (1983:308). I favourably look at this assumption, as in my ontological framework (for which, see later) actual objects are only possible objects which de facto exists. By the way, moreover, this essentialist thesis prevents me from taking another popular objection against direct reference to possibilia into consideration. I mean the semantic objection according to which a DR term designating a possibile cannot but signify what a rigid description means. One may indeed note that, insofar as one considers a description such as "the offspring of a certain Windsor egg and a certain Mountbatten sperm", which allegedly expresses Prince Charles' individual essence, this objection does not specifically concern possibilia, but actual objects as well. As many have noted, however (see for instance Salmon (1982), Recanati (1993:27)), even with actual objects the objection fails to work, be the rigid description in question either a simply de facto one (as that just given) or even a de jure rigid description, such as "the actual offspring of the above gametes". Incidentally, this point is acknowledged by Cook himself; cf. (1983:308).

36. This leaves open the way to a weaker form of the transcendental argument, according to which an imaginative keeping track of an object is necessary in order to directly refer to it. One might find a trace of such an argument in the following remark by Kaplan: "the mere attempt to show that an expression is directly referential requires that it be meaningful to ask of an individual in one circumstance whether and with what properties it exists in another circumstance" (1989a:504).

37. Cf. Quine (1953).

previous page

next page

  contents
 

38. Cf. McMichael (1983:55-6); Plantinga (1979:272). As McMichael (ib.:53-6) comments, actualist theorists who hold such a position must also hold that state of affairs in possible worlds which make sentences containing a proper name "X" true must be existentially generalised state of affairs. This amounts to saying that in the possible world at issue there is no state of affairs containing a possible individual which corresponds to such an existentially generalised state. But as Linsky-Zalta (1994:19-20) note, this position seems hardly plausible. Perhaps it is viable when "X" is associated with a non-rigid description which has no actual denotation, as those considered above. But it is hardly maintainable when the description at stake is an actually unsatisfied de facto rigid description expressing an individual essence. (Perhaps a similar worry is what prompts Corazza (1989:101) to say that reference to possibilia must be meant not in terms of an existential, but rather of a universal, quantification over actual objects. Unfortunately, this latter position cannot in principle work if a description associated with a proper name expresses an (actually unsatisfied) individual essence.)

39. Cf. Kripke (1980:77-8).

40. Cook indeed explicitly acknowledges this. Cf. (1983:308).

41. A different argument against the contraction of possibilia onto unistantiated individual essences is provided by Adams, who says that according to such a thesis one cannot distinguish between possible worlds which differ solely by a permutation of individuals that do not exist in the actual world, since one would precisely need those individuals in order to account for such a difference between worlds (cf. (1977:187-8)). Adams' aim, however, is the negative one of showing that there are neither possibilia nor uninstantiated individual essences. Thus, he could not recognise the validity of the argument provided in the text. Although he acknowleges that actual objects have individual essences, he maintains that these are as contingent as they are (cf. Adams (1981:19-20)). However, we will immediately see that the thesis of the contingency of abstract objects, such as individual essences are, appears to be problematic.

42. I should specify that even descriptions such as "the only explorer resulting from the union of sperm s and egg e", which are not de facto rigid descriptions, since they contain a condition which is both necessary and sufficient for the existence of their denotatum may fix reference of a DR term to it. However, such a fixing is obviously asymetrically dependent on the fixing provided by de facto rigid descriptions such as "the only person resulting from the union of sperm s and egg e". Only if the latter description picks up an individual with respect to a possible world W can the former pick up such an individual as well with respect to W.

43. I here do not commit myself to the thesis that the fixing given by an individual essence-expressing description is sufficient for directly referring to possibilia. If one develops a 'Kantian' conception of possibilia, as I have attempted do to in Voltolini (1991), it may well be that providing proper identity criteria is only one condition for having possibilia at one's disposal, another one being able to think of them intersubjectively, that is, by means of a public language, public names in particular, used to refer to them.

44. Note that the job of fixing reference to possibilia is better done by de jure, or obstinately, rigid descriptions such as "the possible offspring of gametes Z", still expressing a certain individual essence, rather than mere de facto rigid descriptions such as "the offspring of gametes Z". Indeed, by the way it is construed the former description has an actual, albeit actually non-existent, unique denotation; whereas the latter description has only a unique possible denotation. Thus, the former can account for the intuition that a name such as "Oedipus" already actually designates in a direct manner an, albeit unactual, individual. (I borrow the notion of an obstinately rigid description, viz. a description which denotes one individual only in all possible worlds regardless of its existence, from Kaplan (1989b:571).) Now, I can allow for such an intuition if I introduce a slight modification to the

previous page

next page

  contents
 

Kripkean non-realist semantics I would like to espouse: that is, to adopt a 'fixed domain' possibilist semantics instead of the 'variable domain' semantics put forward by Kripke (1971). For a defence of such a possibilist semantics, cf. Zalta (1983). As Tim Williamson has pointed out to me, one might however raise the question as to whether de jure rigid descriptions are already enough to fix reference to possibilia regardless of whether they contain the expression of an individual essence. The interesting case is that of a de jure rigid description which contains less than the expression of an individual essence (the case of a de jure rigid description containing more than an individual essence, such as "the only possible explorer resulting from the union of sperm s and egg e" being immediately treatable as having a reference-fixing feature which is parasitic on that of the corresponding de jure rigid description expressing an individual essence only). Take "the possible son of Philip II and Elisabeth I who is such that a sperm a of his united itself with an Elisabethan egg b". But this description's being modal does not suffice to solve the identity problem we found as regards the corresponding non-modal description. To be sure, even if instead of W we had considered first a further possible world W', in which the relevant individual exists but he is not incestuous, the de jure rigid description above must denote in W' the same individual it denotes in W. But how might we have assessed that this individual is the same as the individual who exists in W' but fails to be incestuous? And if we fail to give an answer to this trans-worldly identity question, how can we assure that there is an individual the de jure rigid description in question picks up (as regards W as much as W')? Once again, therefore, it seems to me that that description picks up an individual conditionally to that individual having been picked up by a de jure rigid description expressing its individual essence only.

45. See, among others, Adams (1981:21). I assume that this objection is not merely terminological - that is, that only propositions having actual individuals as members deserve to be called singular. To be sure, the term "singular proposition" has been evoked by Kaplan to indicate proposition whose members are actual. But, insofar as Kaplan chooses this term to remind the reader of the Russellian origin of the corresponding notion, it would be better to use it to mean propositions whose members are either actual or possible. For early Russell (1937²:43-4,449-50) in order for an individual to be a member of a proposition it does not have to exist, but merely to be (to merely have possible existence).

46. Cf. on this Baldwin (1982); Willamson (1990).

47. It is curious that Plantinga seems to deny that sets whose members exist contingently are necessary subsistent beings, while still holding that propositions involving either individuals or their individual essences would exist in worlds in which such individuals would not and their essences would not be instantiated. Cf. Plantinga (1979:110,121).

48. Cf. e.g. Adams (1981).

49. Modified along the lines suggested in fn.44.

50. Cf. Lewis, e.g. (1979).

51. Cf. fn.38.

52. As Adams (1981:21) explicitly acknowledges.

53. The theoretical possibility of a non-Lewisian actualist Meinongianism similar to that sketched out in the text is manifestly admitted by King (1993:199).

54. For such a position (as well as the attribution of it to early Wittgenstein), cf. Bradley (1989). For the compatibility between a combinatiorialist and a set-theoretical standpoint, cf. Lycan (1993:4,14)

55. A preliminary version of the paper was read at the 1993 Karlovy-Vary (Czech Republic) Symposium on Meaning; I thank all the participants for their comments. Moreover, I am grateful to Roberto Casati, Eros Corazza and Tim Williamson for their insightful remarks to previous drafts of this paper. I would also like to thank Ed Zalta for the fruitful email discussion had with him on these topics.

previous page

  contents