Encoding in Conceivability-Contexts: Zalta’s Theory of Intentionality versus Bourgeois-Gironde’s Notion of Quasi-encoding

1 Introduction

In (Bourgeois-Gironde 2004), the author points out some difficulties affecting Zalta’s Object Theory (OT). His main concern lies with the Modal Axiom of Encoding (MAE) according to which if an abstract object x possibly encodes the property F, then x necessarily encodes F. According to Bourgeois-Gironde, MAE does not account for our intentional relations with abstract objects in conceivability-contexts (such as inquiry and discovery contexts, for instance). More specifically, given MAE, the change in the identity of abstract objects occurs “too soon”. That is, if someone wrongly thinks that an object x encodes the property F, then what they are thinking about cannot be x. For instance, if I erroneously think that Sherlock Holmes lives at 130 Baker Street (instead of 221B), then what I am thinking about cannot be Sherlock Holmes. Bourgeois-Gironde generalizes this fact to the treatment of counteressential situations concerning ordinary objects: for instance, if someone believes that water is not H₂O, they are therefore not referring to water. The author then proposes to replace MAE with “quasi-encoding” which would allow abstract and ordinary objects to possibly encode some properties. Holmes would possibly encode “living at 130 Baker Street”; water would possibly encode “being different from H₂O”.

In this paper, I argue that Bourgeois-Gironde’s concern is misguided. Indeed, OT successfully accounts for the behavior of both ordinary and abstract objects in conceivability-contexts by modeling Husserl’s notion of Noemata. To show this, I will employ some suggestions from Zalta (1988) in which the author sketches a theory of intentionality. Once I have shown that the Noemata-strategy and Bourgeois-Gironde’s quasi-encoding notion are competitors for the same philosophical work, I will argue that the former is superior to the latter. To anticipate a little, the Noemata-strategy allows for a more appropriate treatment of intentionality, while quasi-encoding unduly leaves the burden of all our hypothetical stipulations on the objects to which we refer. Roughly speaking, the Noemata-strategy accounts for the previous cases as follows: Holmes cannot possibly encodes “living at 130 Baker Street”, but his Noema can. Similarly, water cannot possibly encodes “being different from H₂O”, but its Noema can.

As I said, Zalta’s theory of Noemata is only sketched. It therefore needs to be further extended and explored. In this paper, I will then not only show that the Noemata-strategy is in a better position than Bourgeois-Gironde’s, but also – and mostly – that OT can be developed in such a way that it becomes able to successfully account for abstract and ordinary objects in conceivability-contexts. Given Bourgeois-Gironde’s concerns regarding hypothetical stipulations,particular attention will be paid to counterfactual and counteressential statements. In the last part of the paper, I will present issues which the Noemata-strategy faces. More specifically, I will discuss a revenge problem that the supporters of Bourgeois-Gironde’s account could contend, as well as some other difficulties related to reference-issues and abstract objects in different theories.

It is worth noting that the primitive notions of OT are not uncontroversial. The main difficulty concerns Zalta’s distinction between encoding and exemplifying modes which has recently been considered obscure and ad hoc by many commentators.^[1] I think that it can – and deserves to – be defended. However, such an issue goes beyond the scope of this paper. I will merely try is to provide some clarifications of Zalta’s rationale which are indispensable for challenging Bourgeois-Gironde’s idea. Nevertheless, a survey of the Noemata-strategy will incidentally help us to understand the sense of Zalta’s distinction as well as to test the efficacy of another more specific tool associated with OT, namely the treatment of abstract objects that might have been concrete. Such a treatment is presented in Bueno and Zalta (2017) and has to be considered as a reaction to some objections stated in Berto (2013).

Let me now briefly present what I will not treat. Intentionality is, of course, one of the most explored phenomena in modern and contemporary philosophy. The extension of such a phenomenon makes for an interesting debate. In particular, the idea that every aspect of human mental experiences exhibits intentionality can be challenged. Tension between physicalism and intentionality can also be pointed out. These debates are fruitful and need to be addressed. However, this is not the right place. As a starting point, I will consider that at least some phenomena exhibit intentionality in the classical and plausible sense by which they are mental states directed towards objects.^[2]

The paper has 7 Sections. In Section 2, I will present the main formal and non-formal features of OT as described in Zalta (1983) and Zalta (1988). While formulating OT, I will include a presentation of Zalta’s favorite modal framework defended in Linsky and Zalta (1994) as well as his essentialist account proposed in Zalta (2006). This extended formulation of OT will be useful to gain a better understanding of his idea, and so to grasp how to integrate it into a theory of intentionality. In Section 3, I will present Bourgeois-Gironde’s concern with MAE and his reply. Section 4 will present skeptical arguments directed towards Bourgeois-Gironde’s interpretation of MAE. In Section 5, I will formulate Zalta’s theory of intentionality. Then, I will show, in Section 6, how such a theory accounts for Bourgeois-Gironde’s cases. In Section 7, I will present difficulties faced by the Noemata-strategy, arguing that it nonetheless remains superior to Bourgeois-Gironde’s. Section 8 will conclude.

2 Zalta’s Object Theory (OT)

The main aim of OT is to solve some problems affecting Meinong’s theory of objects.^[3] OT quantifies over two domains: that of objects and that of n-place relations. The domain of objects is exhaustively divided into two mutually exclusive subdomains: that of abstract objects (such as numbers, sets, ideas, Sherlock Holmes, etc.), and that of ordinary ones (such as Socrates, my mother, my laptop, etc.). Ordinary objects can be concrete or non-concrete. The meaning of concreteness is given by Linsky and Zalta (1994). Their actualist interpretation of Quantified Modal Logic sets up a fixed domain of objects: ordinary and abstract objects exist at every world. Existence is nonetheless distinct from concreteness: the latter requires spatiotemporal locations, while the former does not. That is, the theorem that states that “everything necessarily exists”, i.e., ∀x□∃y(y = x), is taken to involve an existentially loaded quantifier with no spatiotemporal connotations. Hence, ordinary objects exist at every possible world, but cannot be concrete at all of them.^[4]

Given such a framework, Zalta defines the property of “being ordinary” (O!) and “being abstract” (A!) by using a second-order modal language:

O! = _def λx♢E!x

A! = _def λx¬♢E!x

where E!x means “x is concrete”.

Ordinary objects exemplify properties (for example, my laptop exemplifies the property of “being grey”, and I exemplify the property of “being hungry”). Unlike ordinary objects, abstract ones do not only exemplify, but also encode properties. An abstract object x encodes F – (formally, xF) – iff F determines/characterizes x; by contrast, an abstract object exemplifies x – (formally, Fx) – iff F does not determine/characterize x.^[5] Following Zalta’s examples, the empty set encodes all the properties assigned to it in ZF (such as “being a set with no members”, “being a subset of all other sets”, etc.) and exemplifies the properties that we do not find in ZF (such as “being abstract”, “not having a mass”, etc.); Holmes encodes all the properties assigned to him in Doyle’s novel (such as “being a detective”, “living in London” etc.), and exemplifies properties that do not appear in the novel (such as “being fictional”, “being admired by modern criminologists”, etc.). In general, abstract objects necessarily exemplify the negation of all concreteness-entailing properties (such as “being red”, “being human”, and so on),^[6] and contingently exemplify some converse intentional properties (such as “being imagined by George Clooney”, “being loved by me”, etc.).^[7]

By OT, the particular mathematical theories/fictional stories are analyzed as abstract entities that encode propositional content. The relation between OT and a particular theory T / or a particular story S is given by the Theoretical Identification Principle that assert that the object K of a theory / T or a story S is the abstract object which only encodes the properties Fs exemplified by Kt/Ks according to the theory T/story S, i.e., K _T /K _S = ιxA!x ∧∀F(xF ≡ T/S)⊧FK _T / _S . From this principle, Zalta derives the Equivalence Theorem that says that mathematical/fictional objects encode all and only the properties according to their governing mathematical theory/story, i.e., K _T /K _S F ≡ T/S⊧FK _T / _S . In addition, the encoded properties of abstract objects can be relevantly entailed by those explicitly presented. For example, even though it is not explicitly stated in Doyle’s novel that Holmes is, for instance, a human being, that property is relevantly entailed by, for instance, the encoded property of “being a detective”.

Two more things about the encoding mode. First, an abstract object is complete regarding its exemplified properties but notoriously incomplete regarding its encoded ones. For instance, Holmes does not encode the property of “having a mole on one’s left foot” nor its negation since such a property does not appear in the novel at all. By contrast, for ordinary objects (which encode no properties) the exemplification is always complete. Second, abstract objects do not “lose” their encoded properties from one world to another; so, the encoded properties are rigidly encoded, i.e. ♢xF → □xF. This last point is the Modal Axiom of Encoding (MAE) which will be central to our discussion. MAE is strictly linked with the principle of identity for abstract objects: two abstract objects are identical iff they encode the same properties. Formally, x = _a y = _df A!x ∧ A!y ∧∀F(xF ↔ yF).^[8] For ordinary objects, what counts is the exemplification: two ordinary objects are identical iff they exemplify the same properties. Formally, x = _o y = _df O!x ∧ O!y ∧∀F(Fx ↔ Fy).

Moreover, Zalta’s modal and theoretical framework allows for an account of essentiality (Zalta 2006). Encoded properties play a salient role: they form the set of essential and necessary properties of abstract objects. For example, Holmes necessarily/essentially is a detective who necessarily/essentially lives on Baker Street. By contrast, essential properties for ordinary objects are those that they exemplify at all and only the worlds at which they are concrete. For instance, Socrates is essentially human because he is human only at the worlds at which he is concrete (indeed, “being human” entails “being concrete”); by contrast, Socrates is not essentially self-identical because he is plausibly self-identical at all the worlds at which he exists, and not only at those at which he is concrete.^[9]

Now that we have reached a sufficiently detailed overview of Zalta’s framework, we may proceed with investigating Bourgeois-Gironde’s idea.

3 Bourgeois-Gironde’s Proposal

In Bourgeois-Gironde (2004), the author proposes two interpretations of MAE. The first simply states that abstract objects essentially possess their properties. The second is more complex; it claims that the antecedent of MAE suggests that the existence of abstract objects depends on acts of stipulation. In this way, “x possibly encodes F” means that we can “create” an object which encodes F. Then, given the consequent of MAE, that object would necessarily encode F. Such an interpretation links the creativeness of the agent with the metaphysical claim expressed by the first reading by which abstract objects do not change their properties across the worlds. Using Bourgeois-Gironde’s formulation,

If we bear in mind the fact that abstract objects depend for their existence on acts of stipulation – as the very notion of encoding inclines to think – we can make the antecedent reflect this constructive aspect of our intentional relation to abstract objects. Every act of stipulation deemed acceptable essentially defines an abstract object. What the Modal Axiom of Encoding intuitively means, then, is: if we envision the possibility of an abstract object encoding a certain property, then we have essentially characterized this abstract object. (Bourgeois-Gironde 2004, p. 6).

At this point, it seems natural to raise a question concerning the notion of possibility for encoding. In what sense do abstract objects possibly encode properties? Bourgeois-Gironde proposes two answers: either it means that the encoded proprieties in the antecedent are consistent with other properties encoded by the same object, or, as we suggested above, that the acts of stipulation are “non-empty”: they target property-encoding abstract objects. Bourgeois-Gironde argues for the second answer and points out a problem with MAE. His concern runs as follows:

If each possible encoding of a property to an object freshly individuates a new denizen of the realm of abstracts, there is no way to express something counterfactual about some previously individuated abstract objects that we wish to keep in mind. The problem is clearly that with abstract objects counteressentiality comes too soon – every act of encoding about an intended abstract object shifts the identity of what we are thinking about. Creativeness entails systematic shiftiness. (Bourgeois-Gironde 2004, p. 7).

Clarifications are in order. By “creativeness”, the author means that our intentional states are specified by the content of our act of conceiving; “shiftiness” means instead that the intended object of our thought has been modified and that another one has taken its place. Bourgeois-Gironde’s case for abstract objects is that of a given straight line which plausibly remains self-identical even though it encodes distinct properties when we alternatively consider it in Euclidean and Lobachevskian spaces. Other paradigmatic cases arise in contexts of hypothetical stipulations in such a way that we ignore whether a property belongs to an abstract object. The sequel of the story can take three directions: we rightly predicate an essential property for that object; we wrongly predicate it; the doubt remains. If we rightly predicate it, no shift arises and everything is fine. If, on the contrary, we wrongly predicate it, then the shift occurs. If we do not decide, the identity of the object in question is out of reach.

According to Bourgeois-Gironde, the latter two cases are unintuitive: when we wrongly predicate certain properties for a given object, or we do not know if that object satisfies certain properties, what we “keep in mind” precisely is that very object, and not something else. He then envisages a solution for avoiding the unwanted shift: such counteressential properties are possibly but not necessarily encoded. So, in cases of wrong/uncertain predication, no loss of the referent will follow: we just have those very abstract objects quasi-encoding some properties. What Bourgeois-Gironde then points out is that MAE captures the ontological nature of abstract objects, but fails to account for their behavior in contexts such as discovery or inquiry. Such contexts are precisely what vindicate the role of quasi-encoding, according to him.

Similar cases affect ordinary objects. Indeed, “shiftiness” and “creativeness” are not, as Bourgeois-Gironde points out, specific behaviors of abstract objects in conceivability-contexts but arise, more generally, when we negate some essential properties of any kind of object. Based on what we previously saw, Zalta’s cases for ordinary objects are those in which we negate proprieties that they have at all the worlds at which they are concrete. Just consider the well-known Kripkean example of someone who mistakenly takes water to be composed not of two hydrogen atoms plus one oxygen atom, but takes it to have a different composition. Since “being H₂O” plausibly is an essential property of water, the unintuitive outcome is that what they are thinking about is not water, but something different. By imagining a person’s intentional state, Bourgeois-Gironde claims that

She makes as if ordinary water encodes not being H₂O or, plainly, that she considers the state of affairs of ordinary water not being H₂O in abstracto. […] [W]ater possibly quasi-encodes one of its counteressential properties […] These features of quasi-encoding explain why, whereas we mentally strip an ordinary object of its essential properties, this object may remain intentionally self-identical, and how we can feel epistemically entitled to think that we continue to think and conceive about it what we think and conceive. (Bourgeois-Gironde 2004, p. 9).

Of course, quasi-encoding for ordinary objects involves taking a step back from Zalta’s framework: it does not only reject MAE, but also the idea by which no ordinary object encodes any property. These modifications are nonetheless extremely restricted in extension: quasi-encoding only applies to the conceivability-contexts at issue.

4 Bourgeois-Gironde’s Intentional Interpretation of MAE

My first perplexity arises with Bourgeois-Gironde’s interpretation of MAE. Indeed, Zalta supports MAE with metaphysical reasons which leave no place for such an alternative reading. Two emblematic passages, respectively taken from Zalta (1988) and Zalta (2006), show that the first interpretation of MAE (i.e., abstract objects essentially possess their properties) makes more sense than the second (i.e., the existence of abstract objects depends on acts of stipulation).

[T]he properties that an A-object encodes are rigidly encoded. In other words, if an abstract object encodes a property at one world and time, it encodes that property at all worlds and all times. […] [MAE] ensure[s] that the properties an A-object encodes anywhere anytime are all essential to its identity. (Zalta 1988, p. 24).

♢xF → □xF. This captures the idea that the properties an abstract object encodes are rigidly encoded. Since the properties an abstract object encodes make up the nature of that object, this axiom ensures that each abstract object has a nature which doesn’t vary from world to world. (Zalta 2006, p. 668).

Nonetheless, Bourgeois-Gironde insists that the notion of encoding implicitly suggests his second reading. He argues as follows: “If the role of the Axiom were simply to express the fixed extensions of encoded properties across logical space, xF → □xF” would suffice. (Bourgeois-Gironde 2004, fn. 2). I confess to not fully understanding this passage. It seems true to me that there is a difference between MAE and xF → □xF such that Zalta would assume the first and not the second (as he indeed does). But I also think that this difference has nothing to do with acts of stipulation nor conceivability-contexts; it even goes in the opposite direction. In fact, with MAE, we have that for every abstract object x, if x encodes F at a possible world w, then x encodes F at every possible world. In this way, abstract objects will all encode the same properties at every possible world.

xF → □xF speaks instead about some properties encoded by abstract objects at the actual world. It says that if x encodes F at the actual world, then x encodes F at every possible world. However, xF → □xF does not prevent x from encoding F at the actual world and from also encoding, for instance, G at w2, while it does not encode G at the actual world. Example: Given that Holmes is not complete regarding its encoded properties, xF → □xF allows for a situation in which he has, for instance, a French cousin at w2. By contrast, MAE does not only say that the encoded properties are necessary, but also that all possibly encoded properties are necessary, i.e., Holmes has no French cousin at w2. MAE is therefore stronger than xF → □xF.^[10] That is, the sense of the possibility for encoding seems to be that abstract objects all encode the same properties at every possible world. And this comes from the idea that abstract objects essentially encode all the properties that they encode.

Moreover, we have independent reasons to think that none of Bourgeois-Gironde’s previous interpretations of the antecedent of MAE are plausible. If this means that the properties possibly encoded are consistent with other properties that the object in question encodes, how OT accounts for impossible objects (such as the round-square, for instance) remains unexplained. If otherwise, the antecedent of MAE refers to non-empty acts of stipulation, it remains to understand the rationale of Bourgeois-Gironde’s contexts in which “we try to discover some as yet unknown property of a given abstract object (Bourgeois-Gironde 2004, p. 7). The fact that we try to discover some unknown properties shows indeed that we do not – or at least not always – “create” abstract objects. Pace Bourgeois-Gironde, MAE does not express ”creativeness”.^[11]

Bourgeois-Gironde admits that his reading is not indicated by Zalta; I further think that OT seems to avoid it. This is suggested by Zalta’s elucidation of the nature of the principle of identity:

A definition of identity does not specify how we know or determine whether the entities in question are identical […], rather it specifies what it is that we know once we (pretheoretically) determine that the entities are identical or distinct. (Zalta 1988, p. 30).

By slightly modifying Zalta’s passage, we could claim that MAE does not specify how we know which properties a given abstract object has; it rather specifies what it is that we know once we accept that that abstract object encodes some properties at every possible world.

Enough of Bourgeois-Gironde’s interpretation of MAE. Indeed, even if I am right in claiming that we have good reasons to reject it, his main concern remains relevant. How can OT preserve the “self-identity” of ordinary and abstract objects in conceivability-contexts? Or, in other words, how can we account for the intuitive fact that, for instance, my idea of Holmes and of a particular sample of water will still respectively refer to Holmes and to water even when I am wrong about their essential properties? In the next two sections, I will present what Zalta’s reply could be.

5 Zalta’s Account of Intentionality

Zalta’s account of intentionality in Zalta (1988) and Zalta (1998) stems from the attempt at reconciling Mally’s theory of abstract objects and Husserl’s notion of Noemata. The common background is Brentano’s starting point according to which each intentional state is directed towards something (I think about something, she dreams something, and so on). The “content” of intentionality is however different from the object we refer to. Husserl’s canonical example is the perceived tree that is not the tree of the “external world”, but rather the sense of the perception of that tree.^[12] In particular, Husserl posits the so-called Noemata which organize every intentional state “as if” they were of some objects. The “as if” is crucial because it points out a continuity with Brentano’s tradition for which “every mental phenomenon is characterized by […] the inexistence of an object, and what we might call […] direction toward an object” (Brentano 1973, p. 88). The Noemata are then intermediate representational items that objectify some – so to speak – “non-concrete” contents.

Of course, the main problem affecting intentionality arises with cases involving “another level” of non-existence, that is with states directed towards something that does not exist (e.g., I think about Pegasus, he loves Holmes, and so on). The treatment of such cases is what mostly distinguishes Mally’s approach from Husserl’s. Mally follows his teacher Meinong by postulating directly experienced nonexistent objects; by contrast, fictional names fail, in Husserl’s theory, to have denotation (but maintain cognitive meaning thanks to their Noemata).

The lack of denotation is, according to Zalta, Husserl’s “weak spot”. Indeed, semantically speaking, fictional names contribute to the truth-condition of the sentences which include these states. For non-existent objects, having a denotation would simplify the logic of intentional contexts, as well as our understanding of how such sentences have their anaphoric correlations, entailments, and so on. This is the reason why Zalta proposes to add the Neomeinongian component to Husserl’s theory. Also, Mally’s double predication would serve as a technical device accounting for the relation between the Noemata and their corresponding objects. Let me proceed by showing first how OT displays a theory of intentionality for ordinary objects; after that, I will focus on intentionality for abstract objects.

The intentional act directed towards an ordinary object has A-objects (abstract objects) as content. For the previous case, the external tree is an ordinary object which exemplifies properties (for instance, “being partially green” and “being 3 m tall”). A subject’s intensional state directed towards that tree is associated with an A-object which encodes properties that the tree exemplifies. For instance, if I perceived the tree as being partially green and 3 m tall, my Noema directed towards that tree would be an A-object encoding “being partially green” and “being 3 m tall”. Of course, a difference in properties between the Noema and the object remains. Ordinary objects are characterized by complete exemplification (which is, as Santambrogio suggestively stresses, the seal of their reality);^[13] by contrast, the Noemata are notoriously incomplete in terms of the encoding mode (for instance, my Noema of the tree that I saw yesterday does not encode “being old” nor its negation).

The following passage further clarifies how Husserl’s Noemata directed towards ordinary objects can be modeled by OT in more complex epistemic situations:

A situation in which a person has had a single perceptual encounter with an object. Suppose a particular A-object is the content of the person’s mental state during that encounter (it may, for example, encode just the perceptual properties available to the observer from a certain visual perspective). Suppose further that the person acquires no new information about the object. It now seems plausible to suggest that future mental states directed towards this object, whether they be rememberings, imaginings, fears, etc., will be mediated by the A-object in question. […] Of course, if new information about the object is acquired, some other A-object may come to serve as the content of states directed towards this same object. (Zalta 1988, p. 111).

So far, so good. How does intentionality work for abstract objects? Zalta’s idea goes as follows: once an intentional act is directed towards an abstract object, two abstract objects are involved. One corresponds to the content of the state; the other, to the object of the state. For instance, in the case of fictional characters, we have a content linked to the cognitive capacities of the agent which experiences the state, and the object which has properties in the story in which it originates. Similarly to ordinary objects, intentional states are mediated by A-objects that encode properties. Also, the A-object serves as the noematic content for our future mental states directed towards some objects. For example, the idea of Holmes I have today is an A-object that encodes properties he has in Doyle’s novel (such as “being a detective” and “living in London”) and will serve as content for my future mental acts directed towards Doyle’s Holmes.^[14]

There is still a difference between the treatment of abstract and that of ordinary objects. As previously discussed, for intentional acts directed towards abstract objects, the object towards which the act is directed will not only exemplify but also encode properties; by contrast, in the case of ordinary objects, the object towards which the act is directed will only exemplify properties. As a consequence, the Noema and the abstract object towards which it is directed could eventually be the same object, while – given Zalta’s two principles of identity – this can never be the case for acts directed towards ordinary objects.^[15] Zalta does not explicitly envisage cases of identity between Noemata and the abstract objects towards which they are directed, perhaps because they are very remote: the agent would have to bear in mind not only all the encoded properties of a given abstract object, but also all its relevantly entailed properties. For this reason, intentionality often involves Noemata directed towards objects, rather than the objects themselves. This is also made clear in Zalta’s following passage:

The denotation of ‘Pegasus’ may be an object that has a rather large number of properties associated with it. These are the properties that might be featured in a storytelling of the myth. They include properties that are relevantly entailed […] by the propositions described in a storytelling. Even though there is no single uncorrupted version of the myth, a distinction must nevertheless be drawn between Pegasus the mythical character from the content of someone’s intentional state “directed towards Pegasus.” The Noema that is

involved when the sentence “K is thinking about Pegasus” is true may involve far fewer properties than those featured in a complete storytelling. Stories may be very long, whereas our minds have only so much cognitive capacity for storing properties of the characters described. (Zalta 1988, pp. 106–107).

Up to now, we have seen that combining Mally’s Neomeinongianism with Husserl’s ideas, Zalta offers a theory of intentionality that accounts for the content of thought directed towards both ordinary and abstract objects. In the next section, I will analyze how such a theory applies to Bourgeois-Gironde’s cases.

6 OT for Bourgeois-Gironde’s Conceivability-Contexts

The fact that Zalta posits two abstract objects for the noematic acts says something important concerning Bourgeois-Gironde’s issue. In cases like that of Pegasus, we have two abstract objects: Pegasus from the myth created by someone a long time ago, and Pegasus’s Noema created by the agent at the moment in which their intentional act is carried out. Hence, the person does not, as Bourgeois-Gironde seems to presuppose, create the mythical Pegasus; they rather form their own Noema of Pegasus, namely that in virtue of which consciousness relates to the mythical Pegasus. Such a Noema could not only have, as previously mentioned, fewer properties than the mythical Pegasus, but could also involve some “wrong” properties. As Zalta specifies,

A person might get the details of the story wrong, possibly as a result of mishearing the storyteller. Consequently, the state in virtue of which “K is thinking about Pegasus” is true may be characterized by a content involving properties that are not attributed to Pegasus in the myth. Reasons such as these incline us to try to develop a view on which characters of fiction are distinct from the noemata involved in the states directed towards them. (Zalta 1988, p. 107).

I therefore think that Zalta’s reply to Bourgeois-Gironde’s “shiftiness” in conceivability-contexts where we wrongly predicate an essential property to a given abstract object could go as follows: someone’s partially wrong Noema does not change the identity of the abstract object towards which it is directed. The person is still thinking and referring to that very object, though their idea is not completely right; that is, what they have in mind is an A-object which refers to that very abstract object.^[16] A similar solution applies for the cases in which we do not know whether a property is correctly predicated for a given abstract object. Our doubt would not change the identity of the object in question: we will simply have a Noema which is incomplete regarding that property. Going back to our example, my wrong idea of Holmes as living at 130 Baker Street does not entail any shift in my act of thinking about Holmes. In addition, I can think about Holmes even if I do not know whether, for instance, he lives in London or not.

For ordinary objects, the solution is similar. In thinking about an ordinary object, we form an A-object directed towards it. This A-object encodes properties which could (or not) be the properties that the ordinary object in question exemplifies. If I erroneously attribute essential properties to that object, I will just have a wrong idea of it. If I do not know if a given object has a certain essential property, my Noema directed towards that object will be incomplete. In the previous case of someone mistakenly attributing a chemical composition different from H₂O to water, the person would have an A-object directed towards water that encodes properties which water does not exemplify. If, in an another case, the person just does not know the chemical composition of water, their Noema will be incomplete.

Contrary to what Bourgeois-Gironde argues, Zalta can then account for the so-called “counteressential situations” for both ordinary and abstract objects.^[17] We can express something counteressential about a previously individuated abstract object by individuating a Noema which refers to it and encodes some of its properties (or, alternatively, we can directly individuate that very object by all the properties it encodes and relevantly inherits); then, we create a new Noema directed towards the same abstract object, which encodes nonetheless different properties. Once again, the situation is similar for ordinary objects. We have some Noemata directed towards some ordinary objects. Then, if the encoding properties of a given Noema negate some essential properties exemplified by the ordinary object towards which it is directed, it will represent a counteressential situation.

Moreover, the Noemata directed towards ordinary objects can also account for the case of counterfactual situations. For instance, I can imagine Socrates as not being snub-nosed with a Noema directed towards Socrates which encodes the negation of the property “being snub-nosed”. In this case, according to Zalta’s metaphysics with fixed domains, I am referring to Socrates and picking out a property that he has in at least one world in which he is concrete.

7 Noemata versus Quasi-Encoding

At this point, we need to compare Bourgeois-Gironde’s and Zalta’s solutions by evaluating their pros and cons. More specifically, I will consider six different difficulties which the Noemata-strategy faces, arguing that a solution is viable. Meanwhile, I will also show that Bourgeois-Gironde’s treatment is unsatisfactory.

8 Conclusion

In this paper, I tried to show that OT can be developed in such a way that it solves Bourgeois-Gironde’s counterexamples to MAE. I also pointed out aspects that make OT superior to Bourgeois-Gironde’s quasi-encoding notion. My main aim was to show that OT accounts for a theory of intentionality, that such a theory presents difficulties (that can nonetheless be faced), and that its treatment offers, in turn, a solution to problems affecting OT.^[30]