Voltolini, A. (2020). Why the Mark of the Dispositional is not the Mark of the Intentional. The Journal for the Philosophy of Language, Mind and the Arts, 1(1), 19-32.
[Abstract]In this paper, first of all, I will try to show that Crane’s attempt at facing Nes’ criticism of his two original criteria for intentionality (of reference), directedness and aspectual shape, does not work. Hence, in order to dispense with Nes’ counterexample given in terms of dispositions, there is no need to strengthen such criteria by appealing to representationality, Moreover, I will stress that such criteria are perfectly fine when properly meant in mental viz phenomenological terms that appeal to the possible nonexistence and the possible apparent aspectuality of the object of a thought, its intentional object. For once they are so meant, dispositions clearly lack them.
[Citing Place (1996g)]  
Citing Place (1996g) in context (citations start with an asterisk *):
Section 2 Dispositions Do Not Threaten the Traditional Mark of Intentionality
* Recently, Nes (2008) has maintained that such criteria do not provide jointly sufficient conditions of intentionality. For even dispositions satisfy them: they have both directedness and aspectual shape. In Nes’ own example, take the disposition to attract a metal pretzel. Says Nes,

[e]ven if there are no metal pretzels, something may be disposed to attract a metal pretzel. And even if the extension of “metal pretzel” is the same as the extension of “passion for shrimp-flavoured ice-cream”, i.e. the empty set, the true report:
(1) The ball is disposed to attract a metal pretzel
is not equivalent to:
(2) The ball is disposed to attract a passion for shrimp-flavoured ice-cream. (2008, 209; sentence numbering changed)

By paraphrasing Place (1996), Fn 2: Even though for him things are actually more complicated. See the following footnote. one might say that in looking for the mark of the intentional, one has actually found the mark of the dispositional. Yet as Crane himself stresses (2008, 216), there is an easy way for him to rule out the counterexample, which in point of fact was already presented in similar terms by Martin, Pfeifer (1986). If, as Nes actually does, we consider dispositions in terms of their linguistic reports, it is easy to see why such reports do not provide jointly sufficient conditions of intentionality. For, as many people along with Crane himself (2001) have underlined (starting from Kneale 1968 and Searle 1983), the linguistic phenomena that feature such reports, i.e. failure of existential generalization on the one hand and failure of substitutivity salva veritate on the other hand, are no criteria for singling out the genuine linguistic counterparts of intentional states; namely, adequate reports of such states, intentional reports. Instead, they constitute a mark of the more general linguistic phenomenon of intensionality, which affects dispositional reports just as modal or nomic statements, intensional contexts in general. Fn 3: Place (1996) puts forward an intermediate position. For even if he states that the genuine criteria for intentionality are actually the mark of the dispositional, he rules out aspectual shape as contributing, once linguistically conceived, to mark intensionality instead. As a result, comments Crane, it is no surprise that Nes is right in holding that even linguistic refinements of the above features, such as those involving hyperintensionality or what Nes calls Russellian meanings (the contribution to the structured Russellian propositions expressed by the sentences in which the relevant terms figure) (2008, 213), do not work either.
Section 3 Why the Traditional Mark Works
* Let me start from directedness. Appearances notwithstanding, directedness is not the mental fact that there may be no object for a thought. Instead, it is the mental fact that the object of a thought may exist just as may not exist: the possible nonexistence of the intentional object. By contrast, dispositions are not qualified by directedness so meant. Granted, dispositions may be individuated, if not in a metaphysical at least in a weaker epistemic sense, in terms of their possible manifestations. For example, fragility is the capacity for something to be broken, which if it does not metaphysically depend on this possible manifestation, at least it is epistemically identified by means of it. Yet a disposition is such that it may have no object at all with which such a possible manifestation is related. Pace Nes (2008), this is not the same as what would be a proper directedness for dispositions, if there were any (which is not the case); namely, the idea that they may have an existent as well as a nonexistent object. On behalf of the dispositionalist, one may reply that such an object of a disposition is the possible manifestation itself: a possible event is what the disposition is directed upon (Martin, Pfeifer 1986; Place 1996). Fn 6: For a metaphysical, strong, sense of individuation of dispositions in terms of their possible manifestations, cf. e.g. Bird 2007. By specifying what Martin and Pfeifer (1986) maintain, Place (1996) instead claims that a further criterion that contribute to single out dispositionality is the weaker epistemic identification of something in terms of its object; precisely, its possible manifestation. Crane himself flirts with this idea when he says that dispositions are individuated, in a weak, non ontologically committal sense, by their possible manifestations, just as thoughts are individuated, in the very same sense, by intentional objects (forthcoming; 2001, 25-6). Yet not even this weak epistemically individuative sense of directedness captures the sense of directedness that is involved in the criterion for intentionality. For this latter sense is not epistemic, but phenomenological (if not also ontological), as we will see later: it (possibly correctly) looks to one that one’s state is about something independently of whether it actually exists. Yet again, insofar as there may be no object at all the possible manifestation is related with, this possible event is just a generic, not a singular item, as the object an intentional state is directed upon is instead taken to be. This difference is linguistically captured by their distinct kinds of reports, the intentional vs the dispositional reports.
Let me clarify this point by means of examples. Sean Connery may think of Nicola Sturgeon, the present Scottish First Minister that actually exists, just as he may think of Nessie, the alleged Loch Ness monster that actually does not exist. In both cases, there is something, namely Nicola and Nessie respectively, Sean thinks about; yet simply, in the second case, unlike the first case, that very something does not exist. This is linguistically captured not by the idea that a sentence like:
(3) Sean thinks of Nessie (who does not exist)
elicits no existential generalization, as is traditionally said (e.g. Smith, McIntyre 1982; Searle 1983; and even Crane himself 2001), but rather (see Sainsbury 2018, and even Crane himself 2013) by the fact that it elicits a particular, nonexistentially loaded, quantification. Indeed, from (3) one can validly infer:
(4) Hence, there is something, namely Nessie, Sean thinks about (who does not exist).
Clearly enough, the validity of this inference shows that in the above case there is no failure of existential generalization. For what is rather involved is a particular generalization existentially unloaded – existential unloadedness, for short (McGinn 2000, 2004). For it ranges upon an overall domain of individuals independently of whether they exist or not. While in the dispositional case, existential generalization fails tout court. For, to come back to Nes’ example:
(1) The ball is disposed to attract a metal pretzel
(5) Hence, there is a metal pretzel the ball is disposed to attract
is an invalid inference, even if “there is” is given a non-existentially loaded reading. Indeed, there is no metal pretzel, even in an existentially unloaded sense, the ball is disposed to attract. Granted, in extensional contexts the description “a metal pretzel” actually denotes the empty set. Yet in (1) it has a merely possible denotation, but it actually denotes no actually nonexistent item, not even a possible individual. For it is indeterminate what that possible denotation amounts to, as Kaplan (1973, 505-8; 1989, 609) originally stressed by raising the problem of the insufficient specificity for an actually unsatisfied description to single out a certain possible denotatum. Consider a possible world w that contains a metal pretzel (to be attracted by the relevant ball) and a possible world w’ that contains a metal pretzel (to be attracted by such a ball) as well. Are such possible metal pretzels the same thing or not? There is no fact of the matter as to how this question could be answered. As a result, the step from the de dicto reading conveyed by (1) to the de re reading stated by (5) is illegitimate. Clearly enough, in fact, unlike (3) no plausible existential generalization of any sort generalization, not even a particular one existentially unloaded, does come out of the infinitival expression “to attract a metal pretzel” occurring in (1). Fn 11: As Place himself (1996, 104) implicitly acknowledges.
Again, let me rely on examples. Oedipus may entertain a certain thought with respect to a certain intentional object, call it “Jocasta”, yet fail to entertain the same kind of thought with respect to another intentional object, call it “Mummy”, even if both objects may further appear as aspects of the same thing, Jocasta-aka-Mummy. Sticking to reports of objectual intentional states (but the same result would be obtained if one mobilized reports of propositional intentional states), this is linguistically captured by the fact that (pace Freud) the true:
(6) Oedipus craves for Jocasta
is matched by the false:
(7) Oedipus craves for Mummy.
For in such contexts, the ordinarily coreferring names “Jocasta” and “Mummy” respectively refer to different intentional objects that may further appear as aspects of one and the same thing, Jocasta-aka-Mummy. Thus, as Frege (1892) originally captured, in (6)-(7) there is no failure of substitutivity salva veritate. For there is no referential opacity, but just pseudo-opacity (Voltolini 2005). Indeed appearances notwithstanding, the names do not corefer there, for instead they refer to different intentional objects, respectively named there “Jocasta” and “Mummy”. While in the dispositional case, there is such a failure viz proper referential opacity. Suppose one goes back to:
(1) The ball is disposed to attract a metal pretzel.
(2) The ball is disposed to attract a passion for shrimp-flavoured ice-cream.
Granted, in extensional contexts the two descriptions “a metal pretzel” and “a passion for shrimp-flavoured ice-cream” actually codenote the empty set. Yet unlike what happens in (6)-(7), in (1)-(2) such descriptions do not actually denote different (actually nonexistent) objects; they merely have possible denotations of which is indeterminate whether they are identical. Indeed, it is indeterminate whether there is just one attractable metal pretzel across unactual possible worlds as well as whether there is just one passion-attractable shrimp-flavoured ice-cream across such worlds, hence whether they are identical. As a further result, it is indeterminate as well whether to the two infinitival expressions “to attract a metal pretzel” and “to attract a passion for shrimp-flavoured ice-cream” single out different possible events. All in all, the fact that (1) has a certain truth-value is no guarantee for (2), which turns out from the mere substitution of the first description with the actually codenoting (in extensional contexts) second description, to have the same truth-value. Fn 15: Both Martin, Pfeifer (1986) and Place (1996) appeal to a further criterion that traces back to Anscombe (1965), the so-called indeterminacy of the intentional object, in order to again hold that also this criterion contributes to single out dispositionality, not intentionality. Yet the only plausible sense in which the criterion qualifies intentional objects, which is epistemic – namely, the idea that the subject of an intentional state may not know of certain properties whether they are possessed or not by a certain intentional object – does not qualify dispositions.