1 Introduction

Unlike other notions that philosophy has used from its earliest beginnings in formulating its proprietary set of questions (What are things ultimately (made of)? What do we know? What is a good deed, or a good work of art?), the notion of propositional content, like a priori knowledge or possible world, is a technical-theoretical notion that philosophers deploy to provide answers to those questions. All those who indulge in proposition talk assume some core features of the philosophical ‘job description’ for propositions. Here are some core roles that propositions are posited to play:Footnote 1

  1. i.

    Objects/contents of attitudes like belief, speech acts like assertion, and perhaps others in each category;

  2. ii.

    (Partial) meanings of utterances of declarative sentences, and perhaps others;

  3. iii.

    Semantic values for ‘that’-clauses;

  4. iv.

    (Primary?) bearers of truth and falsity, and their modalities: necessity, possibility, probability (subjective or epistemic, objective or metaphysical);

  5. v.

    What is assessed in determining the validity of arguments or—more in general—the rationality of rationalizing explanations.

These are what we shall call the basic explanatory roles of propositions. What do we mean by this? Start from the thought that there are pre-theoretical phenomena to be explained: (i) what we think or say is endowed with content; (ii) utterances of declarative sentences have meanings; (iii) ‘that’-clauses have semantic values; (iv) truth and falsity have bearers; (v) an argument’s validity or invalidity or—more in general—the rationality of rationalizing explanations depends on the relations between the contents of rationalizing and rationalized conditions. Propositions are posited as parts of (best) explanations of these phenomena. In this sense, they are a theoretical development of the folk pretheoretical understanding of propositions, manifest in talk about sameness of content, and in one of the senses of generalizations such as ‘what S said/thought’.Footnote 2

Despite there being a wide agreement on the basic explanatory roles of propositions, lurking behind central debates in the philosophies of language and mind—such as those about the distinction between singular and general propositions and thoughts, the unity problem, disagreement about matters of inclination, the distinctiveness of indexical attitudes, and so on—there is a fundamental disagreement as to whether propositions can—and perhaps should—play other additional explanatory roles in our philosophical enterprises.

Our broader goal in this paper is to bring out this overlooked fact by systematizing a certain kind of disagreement. The disagreement involves two families of approaches which we will dub propositional inflationism and propositional deflationism. Briefly put, propositional inflationism is a family of views that take propositions to be theoretical entities that raise substantive explanatory demands of their own, and expect explanatory payoffs from such demands, beyond anything strictly required for them to play their basic explanatory roles. Propositional deflationism is the negation of propositional inflationism. We are using the labels “inflationism” and “deflationism” relative to the explanatory job that propositions perform. The inflationist and the deflationist agree on the existence of propositions; they disagree on whether the explanatory role of such entities should be inflated by ascribing them features or natures that go beyond those needed for them to perform the basic explanatory roles described above.

To illustrate: to perform their explanatory roles, propositions themselves must determine truth-conditions, and hence be truth-apt and have truth-values. On the one hand, philosophers like Hanks (2015), King (2019) and Soames (2019) take this to establish that propositions themselves have representational features, which stand in need of substantive explanation; and they offer rival accounts of their nature for that purpose. Deflationists like Merricks (2015), on the other hand, reject such demands for substantive explanation. Now, we want to further distinguish bare from substantive deflationism.Footnote 3 Nobody will deny that the representational features (truth-conditions, truth-aptness, truth-value) of the representational vehicles themselves to which propositions are ascribed require substantive explanation. Substantive deflationists think that such explanations also suffice to meet the corresponding demand for propositions. Perhaps this is because they just think of propositions as derivative from the vehicles (say, equivalence-classes thereof); for our purposes, the reason doesn’t matter. In contrast, bare deflationists like Merricks himself disregard even this derivative form of explanation, adopting a primitivist take on the relevant propositional features. Perhaps this is because they don’t want to ground propositions on their vehicles; again, this is irrelevant for our purposes, and we won’t go into it further.Footnote 4

In what remains of the paper we will see the inflationism/deflationism distinction at work by focusing on the case study of the distinction between singular and general propositions. Our goal is twofold: to show that inflationism about the singular/general divide is mistaken, and to argue that a (substantive) deflationary line offers the best prospects. The paper is organised as follows. In §2 we will elaborate on Quine’s revival of the de dicto/de re distinction and Kripke’s rigidity data about proper names to extract two desiderata for an adequate explanation of the distinction between singular and general propositions. In §3 we devote our critical attention to the inflationary account of singular propositions recently advanced by Glick (2018), who provides what we take to be an excellent and fully worked-out version of inflationism regarding the singular/general divide. Glick defends the following (2018: 1049):

Aboutness without properties (AWP)

A proposition p about an object o is singular wrt o iff p is about o, but not merely about o by concerning one or more properties of o.

We will examine in more detail the idea of a proposition itself being about an object, and concerning an object’s properties, to show that (AWP) fails to satisfy the two desiderata. Finally, in section §4 we discuss a family of substantive deflationary views that we call, borrowing the label from Glick (2018, 1060), vehicle-first approaches (VF):

(VF)

A proposition about an object x is singular wrt x iff p is the content of a vehicle containing a referential constituent [r] for x.

Let us make clear from the outset that we will not go into the details of such a view; that would be too complex a task, which would deprive the paper of the focus it has on advocating deflationary accounts. However, we believe that our discussion will contain enough detail to establish that a version of (VF) explains the data for the singular/general distinction better than any account which, like Glick’s, relies on properties of propositions.

Let’s move then to the distinction between singular and general propositions.

2 Two Desiderata for an adequate account of the singular/general distinction

A number of central philosophical debates—such as debates about quantification into attitude ascriptions, singular (de re) thought, Frege’s puzzles, reference, acquaintance—build on two clusters of data that are held constant and used to adjudicate competitor views on each of these topics. The clusters suggest the existence of a significant distinction between singular and general propositions. They are: (I) The distinction between propositional attitudes that, following in the steps of many philosophers since medieval times, Quine (1956) illustrates with the difference between wanting a specific sloop and merely wanting relief from slooplessness. This distinction is suggestive of a natural psychological kind, that of attitudes with a singular content. Strawson’s (1959) reduplication argument that two qualitatively indistinguishable subjects in a symmetrical universe judging that Everest is cold about qualitatively indistinguishable mountains would nonetheless be making different judgments points, we take it, to this very kind. (II) Kripke’s (1980) and Kaplan’s (1989) rigidity data concerning the contents of names and indexicals. After Soames (2002, 251): while it is false that the object that is the tallest mountain is such that it could not have existed and failed to be the tallest mountain (and no other object could have been the tallest mountain), it is true that the object that is Everest is such that it could not have existed and failed to be Everest (and no other object could have been Everest).

Mindful of (I) and (II), take now the following sentences:

  1. (1a)

    The tallest mountain is cold.

  2. (1b)

    The most massive mountain is cold.

  3. (2)

    Everest is cold.

The Quine-Kripke data suggest that, while (1a)–(1b) express general propositions about Everest, (2) expresses a singular proposition about it. We suggest that these data can be theoretically systematised so as to give rise to two desiderata for an adequate explanation of singular propositions. We will here partly follow the recent discussion offered by Glick (2018). More specifically, we will borrow from him the formulations of the two desiderata. However, we should hasten to emphasise that the desiderata themselves can be understood in an inflationary or deflationary spirit. That is to say, one can unpack the desiderata differently depending on whether one will pursue an inflationary or deflationary strategy to tackle the singular/general distinction. So, the main goal of this section is to present and disentangle inflationary and deflationary readings of the desiderata. This brings into sharper focus how important it is to distinguish between inflationism and deflationism.

The first desideratum is this:

Metaphysical Neutrality (MN)

The singular-general distinction is to be made in a way that is [as] neutral [as possible] on the metaphysics of propositions (Glick, 2018, 1043).

We can take the (I) and (II) data to partly support (MN), since such data mention properties of the semantic contents of the relevant beliefs or linguistic expressions. This suggests that the singular/general divide lies either in properties that propositions have qua propositions or in properties of the representational acts or states to which propositions are ascribed, quite irrespective of what a proposition is. Secondly, (MN) is underpinned by the methodologically sound hope to reach an explanation of the singular/general divide which rests as little as possible on views about the hotly debated metaphysical question of what a proposition is.Footnote 5

However, one might worry that (MN) falls prey to an objection. If there is no uniquely correct conception of what a proposition is—perhaps because the notion of a proposition is but a piece of theoretical machinery used differently by philosophers, there being no fact of the matter about its nature—then different conceptions of propositions may well be committed to different accounts of what a singular proposition is.

This, however, relies on the false assumption that deflationists use the theoretical notion of a proposition in different ways. As has emerged previously, all believers in propositions, both inflationists and deflationists, subscribe to the basic explanatory roles detailed in §1 above. Deflationists insist that propositions are meant to play the basic explanatory roles only. Given that (I) and (II) sufficiently motivate that there might be a robust singular/general distinction worth tracing, this warrants looking for an (as) neutral (as possible) account for it, constrained only by the requirement that it is compatible with such roles.

To illustrate the importance of (MN), notice that such a desideratum invites us to put aside a familiar account of what makes a proposition singular (see e.g. Kaplan, 1975, 724, Nelson & Fitch, 2009, Salmon, 1990, 217), offered in the framework of Russellian structured propositions, to wit, that it is o’s occurring as an immediate constituent of p which makes p singular wrt o. On the account that Glick favours, the justification to disregard this proposal lies on the sketched methodological considerations: we would be better off with an account free from the ontological commitments incurred by the Russellian line.Footnote 6 On our own deflationary take, we should be suspicious that incurring such ontological commitments, going well beyond what is strictly required by the explanatory roles of propositions, can provide any real illumination. We’ll come back to these issues in the next two sections.

The second desideratum that we borrow from Glick goes as follows:

Non-Qualitativeness (NQ)

“A proposition p is singular wrt an object o iff p is about o, but not about o merely qua possessor of certain properties” (Glick, 2018, 1039).

We need to distinguish our own deflationary understanding of (NQ) from Glick’s inflationary one. We should indeed ask: in what sense are propositions about objects? Embracing the deflationary perspective we want to ultimately promote, one might wonder whether there is a sufficiently pre-theoretical notion of aboutness for propositions, as opposed to one for the representational eventualities (e.g. attitudes) to which they are ascribed, relative to which a proposition—say, those expressed by (1a) and (1b)—is “about” an object but merely qua possessor of certain properties.

Glick (2018, 1030–1034) provides in support of the pre-theoretical notion of the aboutness of a proposition (as opposed to the aboutness of an attitude), as well as in favour of the pre-theoretical distinction between the aboutness of a singular and of a general proposition which is encapsulated in (NQ), a quotation from Russell (1903, 316). Here it is, discussing sentences of the form ‘the Prime Minister of England is F’:

[A]lthough, as we have seen, Arthur Balfour forms no part of the meaning of this proposition, yet there is a sense in which this proposition is about Mr. Arthur Balfour. For the proposition is certainly about the present Prime Minister of Engalnd; and the present Prime Minister of England is Mr. Arthur Balfour.

However, it is reasonable to worry that, like most philosophers at that time, Russell probably meant by ‘proposition’ something like interpreted representational vehicle. One might also mention in support Boër and Lycan’s (1986, 125, 132) “grade 1” of de re involvement (which coincides with the relevant weak sense of aboutness, illustrated by the contents of (1a) and (1b) vis-à-vis Everest); it is even clearer that Boër and Lycan mean to classify attitudes, i.e., representational eventualities. Hence, it is far from obvious that we can establish the thesis that it is an intrinsic feature of propositions, that is, a feature of propositions qua propositions, that they are about objects, and that they are about objects in accordance with (NQ)—as the inflationary view would have it.

However, we believe that we can make sense of (NQ)—qua theoretical elaboration of the singular attitudes and rigidity data—in the deflationary approach we recommend. It is perfectly compatible with (NQ) that the relevant sense of “aboutness” for propositions be derivative from a more fundamental one applying to representational vehicles. As we will explain further in the next section, this is indeed the way we think (NQ) should be understood.

Even if perhaps only in this way, we take it that competent thinkers might be ascribed a sufficiently clear intuitive notion of aboutness such that not only sentences (1a) and (1b), but also their contents—perhaps only in a derivative, metonymically extended sense—are about Everest, qua possessor of some properties. As will become more apparent in §3 below, many philosophers—including those with a deflationary perspective on propositions—have found acceptable this talk of their aboutness. Thus understood, and in line with (MN) given our take on it, we regard the contrast that (NQ) assumes just as a modest elaboration of intuitions (I) and (II), barely going beyond them and assuming no more theoretical baggage.

In this section we have offered a theoretical elaboration of the intuitive data (I) and (II). The data give rise to two desiderata for an adequate explanation of the singular/general distinction. This puts us in a position to turn to examine accounts of the singular/general distinction. We will begin with examining and rejecting an inflationary approach.

3 Inflationism and non-qualitativeness

It might seem as if the singular attitudes and rigidity data call for an inflationary explanation. One might reason as follows: since the data suggest the existence of singular semantic contents, then it must be possible to explain the singularity of such contents in terms of their semantic features. However prima facie plausible this might be, though, we believe that no semantic property of propositions can successfully accomplish the explanatory task at hand, and also that inflationary proposals create spurious problems of their own. To argue for this point, we will devote our attention to what we take to be the best version so far of an inflationary approach, due to Glick (2018).Footnote 7

To canvass Glick’s own inflationary account, we deem it instructive to start off by focusing on a popular way of capturing what is special about singular proposition that Glick calls Direct Aboutness, (DA):

Direct Aboutness (DA)

A proposition p is singular with respect to an object o iff p is directly about o, not merely about o in virtue of o’s properties.Footnote 8

Merricks (2015, 187) elucidates direct aboutness along the lines suggested above for (NQ): “for it not to be the case that that proposition is about that entity in virtue of that entity’s having this or that property or being ‘given’ in this or that way, and so on”. As he grants (op. cit., 159), accounts along these lines have been advanced by many authors, including Adams (1981), Bach (1987), Cartwright (1998), Fine (2007), Kaplan (1989), Martí (1995), Plantinga (1974) and Recanati (1993). Along these lines, Bach (1987, 14) makes a popular contrast between the “satisfactional” way in which (1a) and (1b) are about Everest, and the “relational” way in which (2) is.

Now, Glick criticizes (DA) as follows: “whenever a proposition (or sentence or thought) is about an object o, it is about o in virtue of o’s properties, so no proposition will count as singular according to (DA). To see this, note that as long as there is some answer to the question ‘In virtue of what is p about o?’, the answer will be in terms of properties of o” (2018, 1038). He illustrates this with the case of Russellian accounts. On such views, a proposition is about an object by virtue of having it as a constituent. Hence, for any alleged singular proposition about o there would be a property P such that the proposition is about o in virtue of o’s instantiating P—namely, the property being a proposition containing o as a constituent. For (DA) to be compatible with the existence of singular propositions, we should hence take Russellian accounts to be wrong. Therefore, Glick (2018, 1039) argues, (DA) doesn’t satisfy (MN).

Unfortunately for the stability of inflationism, however, as we will show below a very similar objection to the one he raises against (DA) can be made against (AWP)—“as long as there is some answer to the question ‘In virtue of what is p about o?’”. The ultimate culprit for such instability lies in the inflationist assumption on which the argument against (DA) rests, an assumption that deflationists reject. Let us first elaborate on this.

Glick acknowledges that his objection to (DA) works only if one adopts an inflationist view of propositions, referencing Merricks: the objection “can be evaded by someone who holds that there is no explanation of how a proposition manages to represent what it represents” (2018, 1038 fn. 9). But he doesn’t appear to be sensitive to the deep worry this raises for him. Glick rejects (DA), but he does not tell us what his objection to a deflationary reading of it would be. In fact, this is our own take on (DA); we are happy to adopt (DA), or the variants thereof provided by other philosophers subscribing to it previously mentioned, as our “account” of the singular/general distinction, when given in terms of features of propositions themselves. However, we want to insist that, as such, it is not intended as providing any sort of deep explanation, no matter what the rhetoric of ‘in virtue of’ might suggest. We just take (DA) and its variants, exactly like (NQ) as indicated above, as a mere summary articulation of the intuitions in our two clusters of data (I) and (II) suggestive of the singular/general distinction, as opposed to providing any kind of explanation for them.

We take Stalnaker’s account of the distinction, in his possible-world-properties framework, to be given in the same deflationary spirit we want to uphold: “what is it, on this picture, to believe a singular proposition—a proposition about an individual as it is in itself? One can correctly and aptly ascribe a de re belief whenever one can give a correct and unambiguous characterization (in a given context) of the way the world is according to the believer by referring to a certain individual and saying that, in the world according to the believer, that individual is a certain way” (2009, 243). This is just meant, we take it, to state the intuitions in (I) and (II) in terms of a deflationary account: what distinguishes singular propositions from general propositions is that the former, unlike the latter, are individuated (picked out) by referring to the very objects they are both about.Footnote 9

It thus seems that the inflationist can only reject (DA) by taking some sort of stand on the metaphysics of propositions, at the very least a stand that rejects deflationary views such as Merricks’s or ours. Significantly, our move here is not to argue that a certain view of what propositions are should be rejected since, for principled reasons, it cannot deliver the right account of the singular/general distinction. The argument rather runs in the opposite direction: to reject a certain account of the singular/general distinction, a metaphysical assumption on propositions is made. As just pointed out, Glick does not offer any reason in favor of this line of reasoning, but one might think that such a line is motivated by an independently plausible commitment to inflationism in general. One might argue that, since inflationism is correct, then we had better reject deflationism about the singular/general distinction too. However, this kind of defense of the inflationist’s rejection of (DA) is very weak, in that it is far from obvious—for reasons we have outlined at the outset—that inflationary proposals are in general compelling.

We have argued that the inflationist’s rejection of (DA) commits them to disregard a deflationary view on the metaphysics of propositions, thereby losing points in terms of metaphysical neutrality.Footnote 10 We now want now to press the instability point to the effect that the objection raised by Glick to an inflationary reading of (DA) can be equally raised against (AWP):

(AWP)

A proposition p about an object o is singular wrt o iff p is about o, but not merely about o by concerning one or more properties of o.

(AWP) predicts that (1a) and (1b) express different general propositions about Everest, whereas (2) expresses a singular proposition about it. The difference between (1a) and (1b) is explained by the fact that the propositions they express are about Everest by concerning different properties of it: the former concerns tallness (of a mountain), whereas the latter massiveness (of a mountain). The difference between (1a) and (2) is that while the former is about Everest by concerning one of its properties (viz. its being the tallest mountain), the latter is about Everest without concerning any of its properties.

Surely, this indicates that (AWP) may be extensionally adequate. Importantly, since the intuitive differences between (1a)–(1b) and (2) are those captured by our clusters of data (I) and (II) at the outset, capturing such a difference might also suggest that (AWP) meets (NQ). Moreover, since (AWP) does not invoke constituency or any other metaphysical notion, one might conclude that it also satisfies (MN). Once more, if one puts as little into “concerning” as the deflationist puts on “in virtue of” in (DA), we are happy to accept all of this: (AWP) might be taken as another acceptable way of capturing the distinction—one, however, as acceptable as (DA), because both, like (NQ), are just to be taken for these purposes as modest theoretical articulations of (I) and (II), barely going beyond them.

The reader will have realized that the difference between (DA) and (AWP) is in fact subtle; the latter replaces ‘about o not merely in virtue of o’s properties’ in the former with ‘about o not merely by concerning one or more properties of o’. For us, this merely reflects the deflationary point that, to the extent that (AWP) does offer an acceptable “account”, this is because it doesn’t say anything more than (DA) does. Glick, of course, cannot take it that way. Crucially, however, he doesn’t offer a definition of what the concerning relation is. He doesn’t want to rely on features of propositional vehicles for serious explanatory matters, given his opposition to (VF), so he leaves the matter at an intuitive level by making an analogy with aboutness. He claims that “as with aboutness, we can get enough of an intuitive grip on concerning that an explanation relying on that notion can be illuminating even without a definition of it” (2018, 1044). Glick also offers a “heuristic”: “Let ‘F’ be a predicate expressing property P and let ‘F-ness’ be a noun referring to P. Then a proposition q concerns P iff q is expressed by a sentence at least one word of which is either ‘F’ or ‘F-ness’.” But he insists that this “is merely a heuristic, because a proposition’s concerning a property need not essentially involve anything linguistic”.

The claim that we have an intuitive grip on concerning gives us pause. It is indeed important to bear in mind that the discussion focuses on propositions, which are theoretical entities we invoke to do a certain job, whose description is given by claims (i)–(v) laid out at the outset and reproduced above. Surely (i)–(v) are a theoretical development of our pretheoretical/intuitive grasp of propositions. But it’s hard to establish whether we have any intuitions about propositions which go beyond that. To illustrate the worry, suppose we follow Glick’s take on (1a), (1b) and consider now properties indicated by so called phi-features—such as being male for ‘he’, being the speaker for ‘I’, or being called Alex for ‘Alex’—which many contemporary semantic theories take to figure in presuppositions lexically triggered by such features (García-Carpintero, 2011). Do the relevant propositions concern them? Does the proposition expressed by (2) concern being called Everest? How are we to tell?

The inflationist’s reliance on intuition is of no help here. Suppose we want to preserve (MN) and avoid taking any stance within the inflationist vs. deflationist debate about propositions. Since both inflationists and deflationists agree on the basic explanatory roles for propositions, we should wonder whether we can have intuitions about propositions as the theoretical entities playing such roles. Thus, we ask the reader to envisage an experiment involving laypersons with no philosophical training who would have to answer the following question: “Does what one says (or the belief that one expresses) with the sentence “Everest is cold” concern being called Everest?”Footnote 11 In all likelihood, the reaction it would elicit from the participants would be a baffled stare. This suggests that answering the question whether (2) concerns being called Everest by appealing to an intuitive grip on a proposition’s concerning an object’s properties is hopeless.

Let us suppose, however, that we do answer this question in the affirmative; that is, suppose that we maintain that (2) concerns Everest’s property of being called Everest. The consequence is that (AWP) would be objectionable exactly along the same lines that Glick invokes against (DA): to wit, no proposition would turn out to be singular after all; even (2) would be about Everest by concerning some of its properties. As a result, if there are accounts of propositions entailing that (2) concerns being called Everest, (AWP) would fail to preserve (MN), because it would have to exclude them. In what follows, we’ll show by way of some examples that this is not merely an abstract possibility.

Before going into this, let us make clear—to forestall misunderstandings—that, given the possibility of offering a substantive deflationary understanding compatible with (NQ), we do believe that an utterance of the sentence “Everest is cold” is intuitively about Everest without concerning or otherwise involving the name of this object. Similarly, we do believe that an utterance of the sentence “I am hungry” is about its subject without concerning or otherwise involving her being the subject. As we made clear at the outset in articulating our clusters of data (I) and (II), we grant that we have intuitions of singularity. Yet, in our view such intuitions are to be properly explained with respect to the representational act of uttering a sentence (or tokening a thought) rather than by means of features of an entity (i.e., a proposition) that we posit and ascribe to such a representational act for theoretical purposes. Understood in ways acceptable to us, and intuitively plausible, (AWP) merely sums up the data, exactly like (DA) and (NQ) and without adding anything explanatorily illuminating at all to what they already provide.

This clarified, let us go back now to our discussion of concerning, in the alternative, substantive way that the inflationist must be taking it. The supporter of (AWP) might take a new tack, by noticing that while we don’t have an intuitive grasp of a proposition’s being about an object by concerning its properties, we do have aboutness intuitions on the aboutness of representational acts, such as utterances and judgements. Take for instance:

  1. (3)

    I’m hungry now.

An utterance of (3) is, intuitively, about myself, hunger, and the time at which the utterance occurs. Take now:

  1. (4)

    John is hungry.

An utterance of (4) is, intuitively, about John, hunger and the time at which the utterance is made, quite independently of how we chose to semantically represent tenses. That is to say, we can state what an utterance of (4) is about independently of the kind of proposition—temporal as opposed to eternal—that tensed sentences express. Similar remarks hold in the case of cognitive acts (e.g. judging that John is hungry). Theorists such as Soames (2017) and Hanks (2015) claim that propositions are types (as opposed to tokens) of mental or linguistic acts. They both call the main such cognitive act predication, even though they understand it in different ways. For Soames, it is a very general one, entertaining, lacking the specific illocutionary commitments of, say, asserting or guessing; for Hanks it is rather one with assertoric commitments.Footnote 12 Now, the advocate of (AWP) might perhaps think that we should be ready to pay some price in terms of metaphysical neutrality by adopting a view of propositions which might help us to explain what concerning is, thereby explaining what a singular proposition is in a way that meets (NQ). Since the act-theoretic conception of propositions holds the potential to afford an intuitive grasp on the concerning and aboutness of propositions, then we’d better adopt it.

On closer inspection, however, such a suggestion is not going to help an inflationary take on (AWP), in that the combination of (AWP) and the act-theoretic conception of propositions in fact fails to satisfy (NQ). Take Frege’s puzzle. Suppose I assert the following two sentences:

  1. (5)

    I am making a mess.

  2. (6)

    NN is making a mess.

These two sentences express distinct but representationally identical propositions. How can the act-theoretic conception account for this?

Soames (2017) maintains that propositions are act-types which consist of an agent’s predicating a property of an object. Importantly, however, a proposition is a complex act which is made out of other sub-acts. In a case of monadic predication constituting the proposition that NN is making a mess, such a proposition is made out of the sub-act of cognizing the object, and the sub-act of cognizing the property. Given this, we can say that the propositions expressed by (1) and (2) are representationally identical since they both predicate the property making a mess of NN. The predications in question share the same sub-act of cognizing the property. Yet, they differ since they are made out of different sub-acts of cognizing the object. Let us follow Speaks (2016) in visually putting the propositions expressed by (5) and (6) as follows:

  1. (5P)

    = pred <making a mess, NN> ⊕Cog1stPW(NN) ⊕cog(making a mess).

  2. (6P)

    = pred <making a mess, NN> ⊕Cog3rPW (NN) ⊕cog(making a mess).

Now, the act-theoretic explanation of why (5P) and (6P) are about NN involves properties of NN: (5P) is about NN in virtue of NN possessing a given property, i.e. the property of being first-personally cognized by the agent. By contrast, (6P) is about NN in virtue of NN possessing a different property, i.e. the property of being third-personally cognized by the agent. Bear in mind that (AWP) says that a proposition about o is singular wrt to o just in case it is about o but not merely about it by concerning some of its properties. Yet, since (5P) and (6P) are about NN exactly by concerning two different properties of him/her, it follows that, according to (AWP), if the act-theoretic conception of proposition were correct, no proposition would be singular. This shows that the supporter of (AWP) can’t appeal to the intuitive aboutness of representational acts and reduce propositions to such acts in order to explain the distinction between singular and general propositions in virtue of features of propositions themselves. On the contrary, we must exclude this view of propositions if we want to preserve the intuitive distinction between (1a) and (2) and account for it by means of (AWP). (AWP) would thus still fail to meet (MN) if we adopted the act-theoretic view of propositions.Footnote 13

A similar point can be made on the basis of the Evans-McDowell object-dependent views of propositions (Evans, 1982; McDowell, 1977), which, unlike more traditional Fregean views, cannot be accused of not aiming to fully account for the singular/general distinction. It seems that, on any reasonable elaboration of those views, such object-dependent propositions are about the objects they depend on by concerning the properties through which such objects are presented in them. Hence, given (AWP), as with the (variation on) the Hanks-Soames view, if the object-dependence view were correct, there wouldn’t be singular propositions. For (AWP) to be an acceptable account, it needs hence to assume that such views are out of the picture. This is methodologically inadequate, in that it amounts to losing (MN).

It must be said that Glick does have objections to the object-dependence account, but this is irrelevant to the previous point. In fact, given that, as we saw above, he also has (MN)-based objections to the Russellian account, and given that he rejects act-theoretic, and object-dependence views of propositions, as well as deflationary views such as Merricks’s and Bealer’s in his argument aginst (DA), we are bound to wonder: What is left? Rather than metaphysically neutral, (AWP) appears to be metaphysically empty, in that no (inflationist or deflationist) view on propositions can be invoked to develop (AWP) and deliver the singular vs. general distinction.

To sum up the argument in this section, we believe that (AWP)—when interpreted in the inflationary way recommended by Glick—falls prey to a Twofold Objection. First, the view suffers from explanatory emptiness: the inflationist doesn’t offer a definition of concerning, but relies on the idea that we have an intuitive grip on a proposition’s concerning an object’s properties. Since we do not have this grasp, (AWP) fails thereby to explain what it takes for a proposition to be singular. Secondly, the view suffers from metaphysical emptiness: the inflationist rules out deflationary accounts of propositions in order to argue against (DA) and Russellian accounts to argue against (DA) and (MN). They are also committed to rejecting inflationist act-theoretic accounts and the object-dependent account of propositions, for when they are married to (AWP) they yield the verdict that no proposition is singular. Thus, no view of what propositions are seems to be compatible with it.

The Twofold Objection suggests to us that trying to account for the distinction between singular and general propositions in terms of properties of propositions themselves leads us nowhere. (VF) offers a much better account, not subject to these doubts. The task of the following section is to argue for this point.

4 Deflationism and vehicle-first approaches

In this section we outline a defence of:

(VF)

A proposition about an object x is singular wrt x iff p is the content of a vehicle containing a referential constituent [r] for x.

(VF) is a substantive deflationary view. It is deflationary in that it holds that a given property of propositions, i.e. their being singular or general, is not to be explained by features of propositions themselves. It is substantive in that it holds that the explanatory demand of tracing the distinction between singular and general propositions is genuine and can be met by looking at the properties of the vehicles, i.e. representational acts or states such as utterances, thoughts, and the like, to which we ascribe propositions. In this sense, the biconditional whereby we express (VF) is explanatory in the right-to-left direction: it is in virtue of the fact that p is a content of a certain vehicle that p is singular wrt x. Thus, (VF) should be taken to explain the data summed up by (DA) and (AWP) when read in the deflationary way we recommended above.

To forestall expectations we will not fulfil, though, we don’t aim in this section to offer a full-fledged articulation of a specific version of (VF). As we indicated at the outset, we find that goal too ambitious for a single paper. We just intend to provide further support for the claim that the deflationary approach to the singular/general distinction offers the best prospects. Thus, while a full articulation of the right (VF)-based account of singular propositions will have to be deferred to further work, we believe that the following discussion will contain enough detail to support our view that the debate on singular propositions should be conducted within the deflationary camp so as to establish which version of deflationism—whether bare or substantive, and which specific version of (VF) within the substantive camp—is the best one.

This understood, in this section we set out to accomplish three tasks. We will first indicate a consideration in favour of (VF). Then we will cite a key feature of (VF): it allows us to preserve the assumption that the singular attitudes and rigidity data track a significant distinction between singular and general propositions while sticking nonetheless to the “Fregean” idea that the individuation of particular mind-independent objects is grounded, at least partly, on general criteria of identity. Finally, we will respond to four criticisms raised by Glick against (VF).

Let us begin with elaborating on (VF) and mentioning in so doing a point that speaks for it. As Glick (2018, 1029) and before him King (2015, 543) rightly emphasize, the notion of singularity we need to employ is a relative one, singularity with respect to an object. Consider:

  1. (7)

    The forty-third US president respects George H. W. Bush.

(7) does express a singular proposition, “but one wants to add: the proposition is singular with respect to George H. W. Bush, but not singular with respect to the forty-third US president” (Glick ibid.). By specifying that the referential constituent [r] is meant to be referential with respect to x, we aim to capture the relativity of the notion of singularity.Footnote 14 Relatedly, (VF) helps us to realize that the question whether certain sentences express the same (singular) proposition or not is underdetermined by the relativity of singularity. Consider these sentences:

  1. (8)

    Kripke exists.

  2. (9)

    Kripke exists and A exists (where “A” directly denotes Kripke’s biological mother).

  3. (10)

    Bob is here.

  4. (11)

    That man who saw John is here.

(VF) enables us to claim that (8)–(9) can express the same singular-relative-to-Kripke proposition, for (8) and (9) feature the same referential constituent-type for the same object x. However, while (9) expresses a proposition which is also singular relative to A, (8) does not. That strikes us as the right prediction.Footnote 15 As for (10)–(11), (VF) has the resources to explain why they can be taken to express the same proposition. (10) and (11) feature two distinct referential constituents, i.e. [Bob] and [That man], for the same object x. A theory of singular thought or language-vehicles might explain how it is that a mental/linguistic name and a mental/linguistic complex demonstrative can pick out the same object x. As announced, the aim of this paper is not to give a theory of singular thought, but the supporter of (VF) can look at various theories of singular thought (e.g. García-Carpintero, 2021b; Bach, 1987; Jeshion, 2010; Palmira, 2018; Recanati, 2012) in order to choose the one that delivers the best account of this phenomenon.Footnote 16

Having thus expanded on (VF), let us turn to emphasise what we take to be one of its key virtues. Led by the will to respect the singular/general distinction that the data (I) and (II) suggest, several philosophers have come to believe that devices of “genuine” reference should be “Millian”; i.e., they should pick out their referents without any descriptive grounding, as Donnellan (1990, 101, fn.) summarily puts it: the function of a genuine name “is simply to refer without any backing of descriptions, without any Millian connotation or Fregean sense”.

Notice, however, that devices of genuine reference are used, in thought and language, to pick out objective, mind-independent entities, as opposed to sense-data or some such objects of true acquaintance. Various philosophers have argued that singling out particular mind-independent objects depends on conditions that identify them (cf. Burge, 2009, 275, 282 fn.).

(VF) is in a position to maintain that (I) and (II) establish the existence of a meaningful singular/general distinction while, at the same time, accepting the point about descriptive-based individuation of particulars. (VF) indeed allows that the states to which singular propositions are ascribed single out, as such states, the relevant objects relative to general conditions—as explained in different ways in the work of Perry (2012); Richard (1990, 2013; Recanati, (2012) and many others, including García-Carpintero (2021b).

On the face of it, it is doubtful that inflationary approaches can deliver the same result. To illustrate our concern briefly, consider the inflationary account of singular thought offered by Hawthorne & Manley (2012). They propose the following principle (2012, 38):

Sufficiency

Believing a singular proposition about an object is sufficient for having a singular thought about it.

This principle is clearly inflationist: singular thought is understood in terms of propositions. Hawthorne and Manley do not define the notion of a singular proposition, but they consider various glosses, such as “[it is] a content of the sort expressed by sentences containing referential expressions”, or “[it is a content] of the sort expressed by open sentences relative to assignments of objects to their variables” (2012, 4, 13). Tellingly though, Hawthorne and Manly conclude their book by raising the worry that their discussion reveals that we lack any independent grip on the notion of referentiality (p. 248). As Glick notices (2018, 1062), this leads to a skeptical attitude towards the existence of a significant singular/general distinction.

However, given its deflationary stance on propositions, (VF) promotes exactly the opposite direction of explanation to the one adopted by Hawthorne and Manley: singular propositions are to be explained in terms of singular thought, and not vice versa. As has emerged previously, this enables us to take (I) and (II) at face value and respect the point about the descriptive-based individuation of the mind-independent objects picked out by mental and linguistic devices of singular reference.

We have so far outlined (VF). In the remainder of this section, we would like to focus on some objections that can be elicited from Glick’s discussion (2018, 1067–8) of (a different version of) (VF):Footnote 17

  1. a)

    To state that referential terms or concepts differ from non-referential terms or concepts in their semantics because they express different contents would be to get things backward, for it requires an independent understanding of referentiality that presupposes the notion of a singular proposition.

  2. b)

    (VF) might commit us to the inexistence of singular propositions that are not the contents of any thoughts or sentences, but we should be neutral on this issue.

  3. c)

    (VF) does not satisfy (NQ).

  4. d)

    (VF) sheds no light on whether singular propositions can exist in the absence of the objects with respect to which they are singular.

Objection (a) is straightforwardly raised from the inflationary perspective. This criticism thus just begs the question against (VF), in that it presupposes that propositional inflationism is the only way to take up the question of what makes a proposition singular (as well as the question of what bestows representational properties on propositions). Moreover, if our criticism of Glick’s inflationary account of singular propositions are on target, then substantive deflationism is likely to be the only explanatorily interesting way to approach such questions. Thus, far from offering the wrong direction of explanation, (VF) should be regarded as offering the only account of singular propositions that is likely to succeed, along the lines we just outlined.

In reply to objection (b), let us note that the substantive deflationist maintains that real explanatory needs about propositions, such as explaining their truth-conditions, their truth or falsity, their singularity or generality, whether truth is correspondence with the facts, and so on, are to be met relative to features of vehicles. This need not imply that there are no inexpressible singular propositions. David Lewis nicely illustrates the right attitude to have towards this issue while commenting on the deflationary theory of propositional truth. Deflationism about truth tells us that all there is to be said about propositional truth is given by biconditionals of the form [the proposition that p is true iff p], for all propositions. However, Lewis observes that since there’s an uncountable infinity of propositions as opposed to a countable infinity of ‘that’-clauses’, since some propositions cannot be named by ‘that’-clauses of finite length, and since some propositions cannot be named at all, “the story of truth cannot be told completely. But there is no need to complete it. The untold part merely repeats the same pattern over and over, so a tiny sample tells us all we need to know” (Lewis, 2001, 603). Similarly, the fact that there are inexpressible propositions merely means that we cannot tell the whole truth about what makes them singular or general. Yet, the untold part follows the same pattern indicated by (VF).Footnote 18

As for objection (c), we have already stressed how the deflationist is happy to accept (NQ)—like (DA) and (AWP)—in her deflationary spirit. Two further remarks are in order. First of all, the question whether (VF) does or does not satisfy (NQ) boils down to the question of whether we can give an account of the singularity of vehicles, e.g. thoughts, to the effect that they are about objects independently of the objects’ properties. Again, various relational and cognitive theories of singular thought are precisely meant to accomplish this task. Secondly, as already indicated, one might also argue that the intuitive data mentioned in Sect. 2, i.e. Quine’s propositional attitudes data and Kripke’s rigidity data, do not support (NQ) for the representational vehicles themselves (see García-Carpintero, 2021b). This is clearly a decision point within (VF) itself whose analysis is better postponed to another separate investigation. In any case, previous remarks suggest that (VF) has various ways to respond to Glick’s third objection.

Finally, Glick urges that (VF) sheds no light on whether singular propositions can exist in the absence of the objects with respect to which they are singular. Yet, insofar as the singularity of propositions is grounded on the singularity of the vehicles to which they are ascribed, and insofar as there are theories of singular thought that account for the singularity of empty thoughts (see e.g. García-Carpintero, 2021b; Palmira, 2018; Recanati, 2012 for various options) e.g. thoughts about Vulcan, Madame Bovary, the number three (assuming that no such objects exist), (VF) affords a way to defend the existence of singular propositions about objects that do not exist. Again we think that this topic can be more illuminatingly approached by discussing features of the representational vehicles themselves, than by theorizing about propositions and their metaphysics, in contrast also with bare deflationists like Merricks (2015). This is why we believe that a bare deflationary take on the singular/general distinction should in the end be rejected: we do believe—as the inflationist does—that there is explanatory work to do as far as both the representation problem and the problem of the distinction between singular and general propositions are concerned. For this reason, we regard burying the explanatory hatchet and surrendering to sheer quietism as a last resort strategy. Luckily, however, there are substantive accounts in both cases, along the lines of (VF) for the singular/general distinction, to be given at the level of representational acts and their vehicles, which make the substantive version of deflationism we have advocated in this paper more attractive than its bare variety.

5 Conclusions

The question of what makes a proposition singular (as opposed to general) can be taken up by adopting an inflationist or deflationist approach. Inflationism maintains that propositions are endowed with singularity in virtue of one or more features that propositions possess. Glick’s (AWP) is an instance of inflationism insofar as, the way Glick takes it, it holds that it is a feature of propositions, viz. their being about an object o but not merely about o by concerning one or more properties of o, which endows propositions with singularity. However, taken in that way, (AWP) is fatally undermined by the Twofold Objection. First, (AWP)’s appeal to the notion of a proposition’s concerning an object’s properties is explanatorily empty, in that we have no intuitive grip on it. Secondly, (AWP) excludes most accounts of propositions.

In the last part of the paper we have sketched out a defence of an alternative deflationary view, i.e. (VF). (VF) is a substantive deflationary view insofar as it denies that the singularity of propositions is to be explained in terms of features of propositions themselves while, at the same time, maintaining that a substantive explanation of the singularity of propositions can be given in terms of the properties of the vehicles they are ascribed to. Thus, we regard the more specific aim of the paper, namely to illuminate the notion of a singular proposition, as accomplished.

As for the more general aim of the paper—to introduce and illustrate the philosophical significance of the distinction between inflationary and deflationary approaches to propositions—we want to emphasise that the deflationary perspective on the singular/general divide is part of a general deflationary approach to various issues revolving around the notion of a proposition, such as the problem of explaining their having truth-conditions and their truth-aptness.

Within the deflationary approach, various questions remain to be answered. First: is it better to be a substantive or a bare deflationist? Secondly: if we go for the substantive deflationary view (VF), which version of (VF) should we adopt? Thirdly: is it coherent to go for a mixed form of deflationism which maintains a substantive stance on the problem of singular propositions and a bare stance on other issues? These questions will be left for future work. For the time being, however, we rest content with having established that, instead of attempting to provide inflationary accounts of propositions, questions about how we should correctly pursue the deflationary approach to propositions are the ones we should focus on.