Answering Existence Questions in the Best Language for Inquiry

Abstract

Folk ontology seems baroque, compared to the austere ontology of many philosophers. Plausibly, the issue comes down to a choice between existence concepts: the folk and the austere philosophers employ different quantifier meanings. This paper aims to clarify and defend this hypothesis and explore its upshots. How do we choose between the alternative existence concepts; is the austere philosophers’ concept better than the folk’s (or the folksy philosophers’) undiscriminating one? I will argue that contrary to what Ted Sider suggests, the austere existence concept and the corresponding austere answers to existence questions do not prevail in the context of “inquiry”, i.e. the context of practically disinterested pursuit of epistemic excellence. Sider’s suggestion relies on a one-sided idea of epistemic excellence, as beliefs’ conforming to the world. Once we recognize another epistemic aim, namely the agent understanding connections between known facts, ordinary objects will be back in the game.

1 Introduction

Folk ontology seems baroque, compared to the austere ontology of many philosophers. The folk are generally happy to say that there are numbers (for example, numbers greater than 10); that money and governments exist; that there are tables and chairs as well as the particles they are made of; that some Mondays are better than others and some holes are bigger than others. Some of these (explicit or implicit) existence claims are more controversial than others among the folk. For example, the claim that there are numbers might be more controversial than the claim that there are tables and chairs. However, at least the existence of ordinary macroscopic objects is not just carelessly or figuratively implied by the folk, but also believed. This already makes folk ontology look baroque in comparison to the ontology of some austere philosophers who adhere to the ontological version of Occam’s razor. When such austere philosophers establish that one can dispense with quantification over some Fs (e.g. tables, numbers), they take this as evidence (although perhaps not as decisive evidence) that Fs are to be expelled from the ontology, or in other words, announced not to exist.¹ The folk, on the most part, seem not to recognize the relevance of this indispensability test for the truth of “Fs exist”: finding out that quantification over Fs is dispensable (that we could speak in different terms instead) does not affect the folk belief that Fs exist. So, who is right about what exists, the lavish folk or the austere philosophers?

Plausibly, the issue comes down to a choice between existence-concepts: the folk and the austere philosophers employ different quantifier-meanings. This paper aims to clarify and defend this hypothesis and explore its upshots. As has emerged in the debate between Hirsch (e.g. 2009) and Sider (2011), even if ontological issues turn on conceptual choice, the choice might not be arbitrary: some candidate concepts might be better or more relevant than others. Sider insists that not all quantifier-meanings (meaning candidates for “exist”, “there is” and similar expressions) are equally good. There is objective quantificational structure in the world, according to Sider, and it is better to use the existence concept that reflects the objective structure than to use any of the concept’s semantic neighbours. Why? Because it is an epistemic aim, constitutive of inquiry, that beliefs conform to the world. And truth is only a part of a belief’s conforming to the world; being cast in notions that reflect the objective structure of the world is the other part.

Sider’s reasoning provides a way for the austere philosophers’ ontology to prevail over folk ontology. The austere philosophers can claim to be speaking the epistemically privileged language with the joint-carving quantifier. In what follows, I will propose a way of blocking this defence of austere ontology and the corresponding rejection of folk ontology. I will block Sider’s move by rejecting the assumption that beliefs’ conforming to the world is the only relevant epistemic aim. More precisely, I will suggest that another relevant aim is the agent’s understanding of the connections between the facts known to her. In the process of trying to acquire joint-carving beliefs about what exists, I will argue, the agent cannot dispense with her non-joint-carving beliefs about what exists. She then needs to reconcile the truth of her joint-carving and non-joint-carving beliefs about what exists. In order to carry out the reconciliatory inquiry, the agent needs to keep speaking and thinking in a language in which her non-joint-carving beliefs about what exists are (also) true.

In section 2, I will set up the problem: the relevant existence questions (the ones that the folk and the austere philosophers appear to disagree about) come down to a choice between the ordinary and the Occamian existence-concept. In section 3, I will reconstruct the Siderian strategy for arguing that the austere philosophers’ answers to these questions are better than the folk answers (or the folksy philosophers’ answers). According to this strategy, the conceptual choice is not arbitrary: we should prefer the joint-carving concept, to meet the epistemic aim of beliefs conforming to the world. In section 4, I will lay out a framework for assessing the epistemic merits of languages and introduce the epistemic aim of understanding connections between facts. In section 5, I will show how satisfying this epistemic aim requires the agent who has both joint-carving and non-joint-carving true beliefs about what exists to keep speaking a language relative to which both kinds of beliefs are true, for reconciliatory inquiry. In section 6, I will address some loose ends.

2 Existence Questions that Turn on Language

I will be concerned with existence questions, the answer to which turns on language in a certain non-trivial way. Rather trivially, all questions (phonetically or syntactically individuated) turn on the language in which we interpret and answer them. We can assess the question “Is the earth round?” relative a possible English-like language, flat-English, in which “round” means flat; and then, the correct answer is negative (since we would be asking whether the earth is flat, and presumably it is not). However, when we individuate the question by content, then we would not be answering the same question as the English “Is the earth round?” in flat-English. In this paper, I am interested in cases where the same question, individuated by content (at a coarse-grained level), can be answered in two contradictory ways, depending on which one of two versions of a general concept, occurring in the question, we opt for. More precisely, I am interested in the question “Do Fs exist?”, as asked in contexts where the answer turns on which version of the general concept ‘exist’ is better or more relevant.

To exemplify how questions can turn on language in the relevant way, let us look at the question “Is the earth round?” again, but now in a particular context, in which the answer turns on conceptual choice. Let us suppose that a teacher asks this question in class. Two smart students, John and Mary, both know that the earth is a sphere squashed at its poles and swollen at the equator. John answers: “No, the earth is not round”; but Mary says: “Yes, it is round”. In further discussion, it is revealed that the concept of round that John has in mind is “perfectly spherical” and the concept of round that Mary has in mind is “approximately spherical”. Here, John and Mary do not just give divergent interpretations to the string of sounds or letters “round”. They have in mind different versions of the same general concept ‘round’. The question “Is the earth round?”, then, turns on conceptual choice, in this context. John and Mary are not just speaking different languages. At a level of generalization, they are answering the same question, as individuated by content; and they give contradictory true answers to the same question.

What is the teacher to say? She might say that John and Mary’s answers are equally good, since both are true, given the version of ‘round’ that each has in mind. Alternatively, the teacher might say that one of the students has a more relevant or otherwise preferable concept of ‘round’ in mind. If she prefers either John’s or Mary’s answer, we can expect the students to press her to explain why that student’s conceptual choice is better – better without qualification or better in that context. The teacher is then required to justify her position in conceptual ethics: a position on which of the candidate concepts ought to be used in this context or in all contexts.² This paper also concerns justifying a position in conceptual ethics, a position on which concept ought to be used in a context; and the relevant context, here, is that of “inquiry”.

“Inquiry” is a crucial term here, yet an unfortunately obscure one. I will indicate what it means with the help of pointers and examples, falling short of proper definition. “Inquiry”, in the relevant sense, is what we do when we inquire into the world without any practical aims in mind, seeking epistemic excellence for its own sake. Inquiry need not take place in an academic context, but it often does. For example, the situation involving the teacher, John, and Mary is plausibly a context of inquiry. The participants are posing and answering questions about the world in isolation from any practical concerns, with no obvious goal other than epistemic improvement for its own sake. Inquiry, then, is the practically disinterested pursuit of epistemic excellence.

One may be suspicious of whether it even makes sense to speak of good conceptual choices in the absence of practical concerns dictating the desiderata for concepts. I will not engage with such suspicions in the paper. I will assume that we do sometimes find ourselves in such a context of inquiry, where some questions turn on conceptual choice (as they do in other contexts as well); and that we can then choose between the answers by rationally discussing the aptness of the competing concepts for the context of inquiry. My focus will be on how to assess the aptness of the competing concepts for the context of inquiry. More specifically, I will be interested in the case where the question that turns on conceptual choice is “Do Fs exist?” and the contending concepts are the ordinary existence concept and the austere philosophers’ existence concept.

Clearly enough, existence questions do not turn on conceptual choice just in virtue of their semantic content; a certain kind of context is required, for this to happen. Importantly, in such a context, the relevant underlying facts must be uncontroversial. For example, for the question “Are there cats?” to turn on the existence concept, it must be uncontroversial in the context that there are particles arranged cat-wise (or something to the same effect). Further, the context that makes a question turn on conceptual choice need not be that of a dispute. One might be sitting at home alone, thinking about whether the earth is round. One might firmly believe that the earth is a sphere squashed at its poles and swollen at the equator, but remain confused as to its roundness, because there are two versions of the concept ‘round’ available (“perfectly spherical” and “approximately spherical”).

Let us leave the issue of which concept of ‘round’ is better in the context of inquiry (although this might be of independent interest) and turn to the cases that interest us, the existence questions that turn on a choice between existence concepts. Suppose that I am sitting at home alone, thinking about whether numbers exist. There seem to be strong reasons supporting the hypothesis that numbers do exist. For example, I firmly believe that there are infinitely many prime numbers, from which I can easily infer that there are numbers, which is just the same as to say that numbers exist.³ On the other hand, I have taken some philosophy courses and I have encountered the idea that if quantification over Fs is dispensable, then Fs should not be admitted into our ontology, which is just the same as to say that Fs do not exist. And as a friend who has taken even more philosophy classes has told me, Field (1980) has shown that quantification over numbers is dispensable. Frustrated, I give up on numbers and go on to think about whether chairs exist. Again, there is good evidence that they do: I am sitting on one right now. On the other hand, I could very well not talk as if chairs existed and still say everything I need to say about the world. For example, instead of saying that I am sitting on a chair, I could say that I am sitting on particles arranged chair-wise. And yet: it is certainly true that there is a chair that I am sitting on, so there is at least one chair, so chairs exist. Having tortured myself with this for a while, it dawns on me that the question turns on the choice between two existence-concepts. The ordinary concept of existence, familiar to me from life outside the academia, is not sensitive to Occam’s razor like the concept of existence I have encountered in philosophy classes.

Let us also give an example of a dispute context where an existence question similarly turns on conceptual choice. In a philosophy class, the teacher asks: “What do you think, do chairs exist?”. The students all agree that particles arranged chair-wise certainly exist. However, some of them are inclined to say that chairs themselves exist as well. After all, the seminar participants are right now confronted with ample evidence of their existence: there are chairs all around the room. Other students object: strictly speaking, chairs do not exist; what we can see in the room are particles arranged chair-wise and there is no need to talk about the chairs themselves existing – let us keep our ontology minimal. What is the teacher to say; is there reason to prefer some students’ conceptual choice regarding “exist” over that of the others?

One may not be convinced, at this point, that ordinary-language existence claims are insensitive to Occam’s razor, in contrast to the austere philosopher’s existence claims. Maybe the commonsensical students in the class are just wrong in ordinary English? Let us take a short detour to address these doubts. I assume that data about native speaker dispositions can be used as evidence for the truth-conditions of the sentences of the language. I also assume that English speakers, in general, are not disposed to consider the dispensability of quantification over Fs relevant for the truth of “Fs exists”. For example, suppose we convince an intelligent enough English speaker that talking as if rocks existed can in principle be dispensed with, in favour of talking about subatomic particles. I predict that this would not affect the speaker’s belief that rocks exist. The information would not even strike the speaker as relevant for the truth of “Rocks exist”. On the other hand, the austere philosophers are disposed to revise their belief that Fs exists when it turns out that quantification over Fs is dispensable. So, the truth of “Fs exist”, in austere philosophical English, seems to depend on whether quantification over Fs is dispensable; but the truth of “Fs exist” in ordinary English does not seem to depend on this.

I need to take another detour as well, to explain how what I say relates to the similar things said by Hirsch (2009), the most well-known recent defender of the quantifier variance thesis, i.e. the idea that different equally good existence-concepts can be employed in speakers’ idiolects. Hirsch subjects certain existence-disputes to deflationist scrutiny, appealing to the principle of interpretive charity: interpreting charitably, each party should take the other to be speaking the truth in its own language. For example, he tries to deflate disputes on composition (are there composite objects or only indivisible particles?) and persistence (do temporal as well as spatial parts of objects exist?). The composition and persistence disputes may or may not be cases of existence questions turning on the choice of an existence concept; but I am in any case not satisfied with Hirsch’s argumentation to the conclusion that they are. One can question the possibility of giving a satisfactory semantics for one’s metaphysical opponent’s language (especially if the opponent’s ontology includes more things than one’s own); or even more pressingly, one can doubt that charity indeed requires the ascription of linguistic deviance in these cases.⁴ However, regardless of how ontologists should interpret each other and whether they are indeed (just) speaking different languages and their dispute is merely verbal, Hirsch’s more important and less disputable claim is that there can be multiple versions of the same general concept of existence, and so the meanings of quantifier-expressions like “exist”, “there is”, “some” can vary, while still having the same general meaning, in virtue of the inferential role and semantic function of the expressions.

Let us return from the detours to the problem situations described above: (1) the somewhat philosophically informed person pondering on the existence of numbers and ordinary things; and (2) some students in a philosophy classroom appealing to the austere existence concept, while others appeal to ordinary-language truths about what exists. In both cases, we have two competing answers to the very same question, individuated by content; and the question turns on conceptual choice. Is it perfectly alright to answer the question either way or are some conceptual choices better than others?

3 Sider’s Appeal to the Aptness of Joint-Carving Concepts for the Context of Inquiry

In an undemanding sense of “right answer”, both the austere answers and the folksy answers to existence questions are right: both answers are true, given a certain existence-concept. However, are the answers equally good? Sider (2011) argues that some conceptual choices are better than others; and that this also applies to the choice between existence-concepts. More generally, we can take him to appeal to what may be called the “best language principle”: when a question turns on conceptual choice (in a context), then the best among the competing answers is the one that interprets the relevant general concept in the question in the contextually most relevant way. In the case at hand, the context in question is “inquiry” (which, again, means the practically disinterested pursuit of epistemic excellence). Further, Sider claims that the most relevant existence concept, in the context of inquiry, is the “joint-carving” one.

In order to make sense of Sider’s proposal – in particular, the “joint-carving” part – one must first be familiar with Sider’s idea that there is objective structure (there are natural joints) in the world. The objective structure can be reflected by concepts, either perfectly or to a lesser extent. The concepts that reflect the objective structure perfectly are “joint-carving”. The idea that some concepts are better than others at reflecting the objective structure of the world is known from Lewis: according to Lewis, some predicates express natural properties and others express more or less “gerrymandered” properties.⁵ Sider extends this idea beyond predicates (such as “green” and “blue” versus “grue” and “bleen”) to quantification: some quantifier-meanings reflect the objective structure of the world better than others. For this to be the case, the objective structure must involve quantificational structure. The existence-concept that perfectly reflects the objective quantificational structure, then, is the joint-carving quantifier.

Sider does not take a firm stand on whether the ordinary English existence-concept is joint-carving or not.⁶ However, he holds that if it turned out that the ordinary quantifier is not joint-carving, we should answer existence questions in Ontologese rather than ordinary English. Ontologese is a language otherwise like ordinary English, but the quantifier-meanings are stipulated to carve at the joints. In other respects, besides joint-carving, the Ontologese quantifier should be as similar as possible, and similar enough, to the ordinary quantifier (Sider 2011: 172).

Why are joint-carving concepts better than their non-joint-carving semantic neighbours, in the context of inquiry? Sider defends a “thesis about epistemic value”, namely that “it’s better to think and speak in joint-carving terms” (Sider 2011: 61); I would add: “in the context of inquiry”. Sider here and elsewhere sometimes talks as if joint-carving concepts were simply better, without any qualification, but this claim is difficult to sustain, so I will consider the more plausible claim that joint-carving concepts are better in the context of inquiry, as specified in this paper (the context of practically disinterested pursuit of epistemic excellence).

The value of joint-carving concepts, according to Sider, has the same source as the value of truth: “beliefs aim to conform to the world”; and “if belief and the world are both structured, belief aims not just at truth, but also at the right structures” (Sider 2011: 62). So, beliefs aim to conform to the world, and being true is just a part of the job; the other part is being cast in joint-carving terms. Beliefs that conform to the world (including, par excellence, beliefs that are both true and cast in joint-carving notions) do not have mere instrumental value; conforming to the world is “a constitutive aim of the practice of forming beliefs” (ibid.: 61). Primarily this case for the value of joint-carving belief allows Sider to claim, at a later point in the book, that all quantifier-meanings are not on a par. Since there is a privileged, joint-carving quantifier-meaning, we should prefer Ontologese, the language with that quantifier-meaning, over ordinary language, if the ordinary quantifier does not carve at the joints.

Recall now the thesis of the previous section: the austere philosophers, in the context of doing metaphysics, employ a somewhat different existence concept from that of the folk and the folksy philosophers, such that the truth of “Fs exist” is sensitive to whether quantification over Fs is dispensable. Now suppose that the austere quantifier is the joint-carving quantifier. Suppose also that Sider is right about the superiority of Ontologese, at least in the context of inquiry. Then the austere philosophers’ answers to existence questions would triumph over the folk answers, in the context of inquiry. The relevant language of assessment for “Do Fs exist?” would not be the ordinary language, but its philosophical modification, because the modified language is better suited for the context of inquiry (the practically disinterested pursuit of epistemic excellence). The answer that chairs exist would come out as the worse answer (supposing, again, that the Occam-quantifier is joint-carving and the ordinary quantifier is not). And this would be the case in any context of inquiry, not just in the philosophy classroom. If I am thinking about whether chairs exist, at home, but in the spirit of inquiry, then I should opt for the joint-carving Occam-quantifier that is better suited for inquiry and draw the conclusion that chairs do not exist.

One way to block this move against folk ontology would be to dispute the claim that the Occam-quantifier carves at the joints, or even to dispute the intelligibility of joint-carving quantification. One might also dispute the best language principle; but I find it plausible that if a question turns on language, in the context of inquiring into the world, we should prefer the language that is more apt for the context of inquiry. I will pick up on another assumption of Sider’s. In the next section, I will show that this kind of defence of austere philosophical ontology over the relatively lavish folk ontology depends on a one-sided conception of the relevant epistemic aims in the context of inquiry. Namely, Sider apparently supposes that all we (should) care about, in the context of inquiry, is beliefs’ conforming to the world. I will additionally introduce the epistemic aim of the agent understanding the connections between the facts known to her. This brings ordinary objects back in the game, even when the ordinary quantifier is not joint-carving.

4 Understanding Connections as a Relevant Epistemic Aim in the Context of Inquiry

The best language principle suggests a systematic approach to answering questions that turn on conceptual choice in a context. For ontologists, the further idea of the context of “inquiry” and the best language for inquiry just might provide the most promising prospect for redeeming ontologists’ ambition to establish what “ultimately” or “really” exists. Fs would “really” exist iff Fs exist relative to the best language for inquiry. However, this strategy calls for some hard work in sorting out the relevant epistemic aims in the context of inquiry.

Recall Sider’s idea that joint-carving concepts (concepts that reflect the objective structure of the world) are more apt than non-joint-carving concepts for the context of inquiry. Joint-carving concepts are supposed to be better for the context of inquiry because they facilitate joint-carving belief; and joint-carving belief is a relevant aim in the context of inquiry. The desirability of joint-carving belief, for inquiry, in turn derives from the desirability of beliefs that conform to the world. However, is beliefs’ conforming to the world the only relevant epistemic aim, in the context of inquiry? I will call this into doubt.

First, however, let us take a step back and ask: how are we to assess the value of a language (understood as a set of concepts) for inquiry, i.e. for the practically disinterested pursuit of epistemic excellence? A natural answer is that a language L1 is more apt for the context of inquiry than a language L2 iff L1 is more conducive to the satisfaction of the relevant epistemic aims than L2. Now, how to assess the satisfaction of the relevant epistemic aims? I suggest that a relevant epistemic aim is satisfied when a relevant epistemic target of evaluation meets a relevant epistemic success-condition. For example, epistemic targets of evaluation include beliefs, belief sets, cognitive systems, agents, and minds. Epistemic success-conditions include conforming to the world, (less demandingly) being true, having knowledge, having understanding, being coherent, being rational, and so on.

Not all target/success-condition pairs are appropriate: for example, a belief or a belief set cannot have knowledge and a mind cannot be true. Epistemic aims, then, do not include knowledgeable beliefs or true minds; but they might include, for example, true beliefs, coherent belief sets, and rational agents/minds. One might also have a broader conception of epistemic aims that only specifies the relevant domain of targets, such as beliefs, belief sets, etc.⁷ Then, John’s epistemic aim might be that his beliefs would make him happy (or Mary’s epistemic aim might be that John’s beliefs would make him happy). The aim would be epistemic just because its satisfaction would require an epistemic target – belief – to be in a certain way. However, on my narrower conception of epistemic aims, which I think is relevant for understanding what is meant by the aims of “inquiry”, the practically disinterested pursuit of epistemic excellence, the satisfaction of epistemic aims requires epistemic targets to be in a certain epistemic way.

I will not attempt to specify the necessary and sufficient conditions for something being an epistemic target of evaluation and some state being its epistemic success-condition. I will rely on our shared grasp of what falls under the label “epistemic” and what does not. Clearly, one may assess belief, an epistemic target, with regard to a non-epistemic success-condition like “making the bearer happy”. Similarly, one can distinguish epistemic evaluation of minds and agents from other sorts of evaluation. For example, rational belief revision is an epistemic success-condition of an agent, whereas acting morally is not. Examples of non-epistemic targets meeting epistemic success-conditions are harder to find, but candidates include the talk of “true love” and “rational organization of society”. True love and rational organization of society are not epistemic aims, on my view, and neither are happiness-conducive beliefs or moral agents.

Sider considers only a certain kind of target of epistemic evaluation, namely belief; and a certain kind of success-condition, namely conforming to the world (which includes truth and joint-carving). I think there is another relevant target of evaluation and its success-condition, in the context of inquiry: the target is the agent and the success-condition is understanding connections between the facts known to the agent. For example, consider the desired effect of a university course on the student: it is not good enough if the student acquires an extensive set of true beliefs on the subject matter; it is desirable that she would also grasp the connections between these facts, and even the connections between these facts and those taught in other courses, thereby developing a systematic, coherent, transparent general picture. I suppose that a student’s failure to grasp connections is considered a failure precisely because it shows that in an important way, the agent has not improved epistemically, as an inquirer (which is what university courses, generally speaking, are supposed to achieve). Regardless of whether understanding connections may also be practically desirable, it is plausible that at least of a part of the problem here is that the student has not improved along a dimension of intrinsically worthwhile epistemic excellence.

The idea that it is desirable to understand connections between facts is also the basis for a popular way for philosophers to establish the identity of their discipline, at a time when most specialized questions seem to be dealt with by other fields. A common solution to the identity crisis is that philosophers are those who “put the picture together”. It is taken to need no further justification that “putting the picture together” needs to be done, as a part of inquiring into the world. Also, explaining how some facts relate to others, and specifically, how non-fundamental facts relate to fundamental facts, is one of the central projects of contemporary (meta)metaphysicians, including Sider himself.⁸ Sider suggests that ordinary claims made in non-joint-carving terms have “metaphysical truth-conditions”, and providing such truth-conditions shows how what we ordinarily say “fits into fundamental reality” (Sider 2011: 112).

Recall that Sider takes beliefs’ conforming to the world to be a constitutive aim of inquiry, not just a worthy epistemic aim that is relevant in the context of inquiry. I am not sure whether this makes a difference to the argument, but I do propose that understanding connections is similarly a constitutive aim of inquiry. It is one of the necessary elements we expect from inquiry; its lack makes us question whether what we are observing inquiry (the practically disinterested pursuit of epistemic excellence) at all. Suppose we met an alien species whose scientific treatises are mere lists of sentences representing individual well-confirmed beliefs and their justification, with no attempts to develop general theories that illuminate the connections between the facts. They consider their best inquirer to be the alien with the most true beliefs. We would be surprised because when one engages in activities like writing scientific treatises, one should care about the connections; it is a part of what inquiry is about. Without it, the pursuit of epistemic excellence, in our sense, remains incomplete. We would be surprised about these aliens almost as much as we would be surprised to find aliens who, in a context that otherwise looks like inquiry into the world, place superior value in beliefs that make them happy (valuing truth only insofar as it is conducive to happiness) and consider the happiest person to be the best inquirer.

The agent’s understanding of the connections, then, seems to be a constitutive aim of what we mean by “inquiry”, in the relevant sense, on a par with beliefs conforming to the world. Just as we have a notion of what it is for a belief to improve towards epistemic excellence – to conform more to the world – we have a notion of what it is for an agent to improve towards epistemic excellence, and it includes, among other things, grasping connections between the facts known to her. We think of truth (in the case of beliefs) and understanding connections (in the case of agents) as obvious epistemic success-conditions that should be strived for in activities of inquiry, such as the writing of scientific treatises, and taken into consideration in bestowing honours like “the best inquirer”.

However, it does not seem to be desirable to understand the connection between any two propositions that we know to be true. Understanding connections is desirable, for example, when there is a (presumed) connection; or when there is an appearance of a contradiction between the facts – when it is not clear how two or more well-confirmed beliefs or belief sets can all be true. For example, it is this sort of appearance of contradiction that makes Searle ask: “How can there be an objective world of money, property, marriage, governments, elections, football games, cocktail parties and law courts in a world that consists entirely of physical particles in fields of force, and in which some of these particles are organized into systems that are conscious biological beasts, such as ourselves?” (Searle 1995: xi-xii). In the next section, I will argue that in the case of beliefs formed using the ordinary quantifier and the joint-carving quantifier, such an appearance of contradiction emerges, and one needs to undertake inquiry into how such different facts fit together into one world. This suggests a new way of defending folk ontology, while acknowledging that the ordinary quantifier is not joint-carving and that the Occam-quantifier might be.

5 Rescuing the Folk Answers to Existence Questions

We saw how Sider’s appeal to the superiority of Ontologese for the context of inquiry can be mobilized against ordinary ontology. The Occam-quantifier might not be exactly what Sider means by the joint-carving quantifier, but the Occam-quantifier, at least, is what many other philosophers seem to have in mind as the superior quantifier. The rationale for this has to do with the idea of parsimony as a theoretical virtue conducive to getting at the joints of the world. The austere philosophers can draw on Sider, then, to insist that the Occam-quantifier is joint-carving and therefore the relevant quantifier for answering existence questions, in the context of inquiry (the practically disinterested pursuit of epistemic excellence).

I will argue, however, that the best language for inquiry, for an agent who holds ordinary-language beliefs about what exists, should also involve the ordinary quantifier-meaning, even when the latter is not joint-carving. The agent who adopts the philosophical quantifier-meaning into her language and tries to speak Ontologese, in the context of inquiring into what exists, will continue to also have true non-joint-carving beliefs about what exists. At a coarse-grained level, the agent’s joint-carving and non-joint-carving beliefs about what exists are contradictory. Yet both sorts of beliefs are true and therefore represent facts (in some appropriately thin sense). This puts pressure on the agent to understand the connection between these facts. And the inquiry into the connections can only be conducted in a language where both the joint-carving and non-joint-carving answers to the relevant existence questions are true. For that to be the case, the language has to include the ordinary-language quantifier as well as the joint-carving quantifier. This is the short version of the argument; now let us look at some of the details.

First, why will the agent continue to have non-joint-carving beliefs about what exists, after adopting the Ontologese quantifier in pursuit of epistemically superior beliefs about what exists? After all, if an agent is to satisfy the epistemic aim of all her beliefs fully conforming to the world, she should presumably drop her belief that numbers and chairs exist. The belief is false if cast in joint-carving terms; and if true, it is cast in non-joint-carving terms. Either way, it fails to fully conform to the world. If the agent’s beliefs are to fully conform to the world, they should all be both true and joint-carving.

In response to this: let us go back to me sitting at home, pondering on whether chairs and numbers exist. I believe that there are, in the ordinary sense, numbers and chairs; and it is a true belief. True beliefs represent facts. So, it is a fact that there are numbers and chairs. How can it be epistemically bad for me to believe a fact? Is it not even epistemically bad if there is a fact that I fail to believe? It should be at least epistemically permissible, if not desirable, for me to continue to hold my true belief that numbers and chairs exist.

Another worry, concerning the requirement to drop the non-joint-carving existential beliefs, is the great practical resistance to this. They are beliefs that I have had for a long time and I encounter confirming evidence for them all the time. In the case of chairs, for example, I see (or otherwise perceive) them every day, most of the day, and it would be hard for me not to re-acquire the belief that they exist, even if I supressed it for a while. Further, I need to use the ordinary existence-concept in daily communication, and it would be hard to keep reminding myself to only use it in a pretenseful, “as if” spirit and not really believe what I say when I say that there are chairs.

A further, possibly most important worry is that the ordinary quantifier seems to be prior in one’s reasoning to the Occam-quantifier, in the following way. In order to establish what exists_Occam, I have to start out with ordinary truths about what exists and work my way to what exists_Occam by a process of reduction into progressively more fundamental terms – what exists, metaphysically, is what remains after I have left out of my ontology everything I can dispense with. If I vigorously try to forget about the ordinary existence concept, in the context of inquiry, I seem to lack the necessary data for figuring out what metaphysically exists. I can only acquire true beliefs about what exists, in the austere metaphysical sense, by first endorsing my true beliefs about what exists, in the ordinary sense, and then reductively paraphrasing as much as I can.

Given these considerations, it seems that when I make it my goal to acquire joint-carving beliefs about what exists, I will continue to also have my old, ordinary and non-joint-carving beliefs about what exists. This puts pressure on me to recognize the facts represented by all those beliefs as facts known to me. Further, since there are apparent contradictions between these facts known to me, I am pressured to inquire into their connections, in the context of inquiry.

Granted, at a fine-grained level, my beliefs about what exists would not be contradictory, because they involve different existence-concepts. However, remember that exist_folk and exist_Occam are versions of the same general concept ‘exist’. That is why we can answer the same question about whether Fs exist (individuated by content) and get either an affirmative or negative answer, depending on which concept we opt for. Now, if an agent has both joint-carving and non-joint-carving beliefs about what exists, and at a coarse-grained level, these beliefs are contradictory, we can expect this to arouse a certain uneasiness in the agent, the sort of uneasiness that calls for inquiry into the connections between the facts that these beliefs represent. It is that contradiction at a coarse-grained level, the possibility of answering the question with the same content affirmatively or negatively, that gives rise to the need to understand how the beliefs can both be true, how the facts they represent – facts known to the agent – can both be the case. In order to inquire into the connections between the two sorts of existential facts, the agent should speak a language that includes the ordinary quantifier-meaning as well as the metaphysical, joint-carving quantifier-meaning.

One may respond that when doing philosophical ontology, only certain kinds of epistemic aims are relevant. The ontologists’ conceptual choices therefore need not be the best choices for inquiry as such, but for this sort of inquiry, i.e. ontological inquiry. Consequently, the language spoken in the ontology room just needs to be the best language for ontology. One way of putting this is that metaphysicians are asking existence questions, presupposing a limited set of interests that contextually restrict the ordinary quantifier in the ontology room. The agent’s aim of understanding the connections between all the existential facts known to her just would not be relevant in the ontology room; she can conduct this reconciliatory inquiry elsewhere. However, this puts on the metaphysician the burden of explaining the special purposes of philosophical inquiry into what exists; and this is known to be a question that ontological realists would rather avoid. Further, the philosophical answers to existence questions would then be the best answers only relative to a limited range of aims that may well seem quite arbitrary to outsiders. Metaphysicians could not claim themselves to be establishing what ultimately or really exists, with any plausibility. The ultimate answers, if there are any, would presumably be true in the language that best satisfies all the general aims of inquiry (understood as the practically disinterested pursuit of epistemic excellence), including the aim of understanding connections between known facts.

Finally, a note on how the argument advanced here improves the situation for the folksy philosophers (the philosophical proponents of folk ontology). The folksy philosophers have not done much to dispute Sider’s resort to Ontologese, the “better language for inquiry”. For example, Schaffer discusses Field’s (1980) denial of the existence of numbers on the grounds of dispensability, and suggests instead construing Field’s proposal as showing “how numbers exist in a world of concrete substances” (Schaffer 2009: 360). This would be preferable to drawing the conclusion that numbers do not exist “because it reconciles Field’s view with the obvious fact that there are prime numbers” (ibid.). But if we speak Ontologese, the latter is not an obvious fact, it is merely a fact in flawed English; and Schaffer does not explain why it is desirable to keep speaking a language relative to which it is an obvious fact, in the context of inquiry. To counteract the Siderian “inferior language” attack on plain English ontology, I propose picking up on Sider’s assumption that Ontologese is the superior language for inquiry. Sider cannot simply block my recognition of the facts of folk ontology by claiming the non-joint-carvingness of ordinary English. He must instead explain why the epistemic aim of understanding connections between the facts known to the agent is not relevant in the ontology room. I have thus suggested a way for the folksy philosophers to hold their ground against the austere philosophers equipped with Sider’s weapons, without needing to deny the intelligibility of joint-carving quantification, nor our ability to introduce a joint-carving quantifier, nor that there are quantificational joints in nature, nor that the joint-carving quantifier is the Occam-quantifier.

6 Loose Ends: Implications for the Truly Baroque Ontologies and whether the Epistemic Aims Are Jointly Satisfiable

Before wrapping up, I will address a couple of general questions about what my position implies or assumes.⁹ First, I have only considered how we should answer existence questions about “ordinary objects”, i.e. those things that the folk readily admit to exist. However, another kind of existence question causes an almost equal amount of trouble in contemporary ontology: it is the question about whether certain “extraordinary objects” exist. Ontologists argue not just about whether there are chairs and tables, but also about whether there are, for example, incars and outcars. An incar is the part of the car that is inside the garage (or the whole car, if the whole car is inside the garage) and an outcar is the part of the car that is outside the garage (or the whole car, if the whole car is outside the garage). As a car drives out of the garage, the incar gets smaller until it disappears, and the outcar first appears and then gets bigger, until reaching the size of a full car. The example is due to Hirsch (1982: 32).

It is widely believed that the folk side with the austere philosophers in denying that there are incars and outcars (or similarly contrived objects) – although the folk ontology is more baroque than the austere philosophers’ ontology, it is not as baroque as to include such odd things. But there are, of course, the ontologically promiscuous philosophers, or the “baroque philosophers” as we may call them, who answer affirmatively to the existence questions about such extraordinary objects. And my reasoning thus far may seem to imply that the baroque philosophers’ answers to existence questions about such extraordinary objects should prevail, in the context of inquiry, over the folk’s and the folksy philosophers’ answers. Then I would not have succeeded in defending folk ontology, after all.

Why would the baroque answers to existence questions about extraordinary objects prevail, in the context of inquiry? One might say that in the baroque philosophers’ language, sentences like “Incars exist” are true and thus they represent facts (in the relevant thin sense). Thus, the baroque philosophers, the folksy philosophers and the austere philosophers and all others in the context of inquiry should recognize such facts as facts known to them. And once such facts are recognized, all the inquirers (the austere, folksy and baroque philosophers alike) need to reconcile them with the other known facts, that incars do not exist_Occam or exist_folk, given the coarse-grained contradictions between these facts.

However, I think that the baroque philosophers cannot make use of my reasoning to vindicate their position, in the manner outlined. The main reason is that it is not plausible that we should recognize as “facts known to us” the facts of all possible languages that we could conceivably speak or even those that someone that we happen to interact with speaks. So, it is unclear why the folksy philosophers or the austere philosophers should recognize the fact that there are incars as a fact known to them. Further, it is not even clear whether the baroque philosophers themselves should recognize it as a fact known to them. There is an important difference between the fact that there are incars and the fact that there are chairs: the difference is in how badly inquirers need to recognize these as facts known to them. The austere, folksy and baroque philosophers all badly need to recognize the fact that there are chairs, as a fact known to them. There are significant practical pressures for this: we are required to think and speak in such terms in everyday contexts and this pressures us all to keep forming beliefs cast in these concepts. The austere philosophers additionally have the theoretical pressure that their procedure for acquiring true beliefs about what exists_Occam requires first forming true beliefs about what exists_folk. As long as we cannot identify a similar practical or theoretical need to form beliefs using exist_baroque, we are not under similar pressure to recognize the facts represented by such beliefs as facts known to us that need to be reconciled with the other, apparently contradictory facts.

Another issue that might be raised is that I have spoken so far as if the aim of understanding connections can be simply added on to the aim of beliefs’ conforming to the world (“the aim of conformity”), in the context of inquiry. However, there is reason to suspect that the aim of understanding connections interferes with the full satisfaction of the aim of conformity. Conformity has two aspects: concepts carving at all the joints that there are, and concepts not carving where there is no joint. Regarding the first aspect of the aim of conformity, there is no conflict with the aim of understanding connections. The second aspect of the aim of conformity can come into conflict with the aim of understanding connections, however, when there is practical or theoretical pressure to recognize non-joint-carving facts as facts known to us, and there is also some pressure to understand the connections between these non-joint-carving facts and the joint-carving facts (e.g. arising from the appearance of contradiction). My view is that in such cases, the aim of conformity (in particular, the aspect of not carving where there is no joint) must indeed be sacrificed for the sake of understanding connections; but I will not make a case for that conclusion here.

7 Conclusion

Sometimes, questions turn on conceptual choice. This happens with existence questions as well: our answer to “Do cats exist?” sometimes depends on whether we opt for the ordinary existence-concept or the austere Occamian existence-concept, while all the underlying facts (such as the existence of particles arranged cat-wise) are uncontroversial. Sider has suggested a way to banish the ordinary-language truths about what exists from the inquiry into what really exists: Ontologese, the language with the joint-carving quantifier, is superior to ordinary language, in the pursuit of epistemic excellence. Ontologese is superior because joint-carving beliefs are better at conforming to the world than merely true beliefs; and beliefs’ conforming to the world is a constitutive aim of inquiry.

I argued that another relevant epistemic aim in the context of inquiry can be appealed to, in order to keep the ordinary quantifier in our language of inquiry. That aim is understanding connections between the facts known to the agent. The agent inquiring into what exists, in the joint-carving sense, starts out with beliefs about what exists, in the ordinary sense. These beliefs cannot be easily dispensed with, and it is not clear whether they should be. First, the beliefs are true and therefore represent facts and it seems at least permissible if not desirable to sustain such beliefs. Second, there is practical pressure to sustain the non-joint-carving beliefs about what exists: there is a constant inflow of confirming evidence for the beliefs; and having the beliefs facilitates communication with others who use the ordinary quantifier. Thirdly, the procedure of establishing what exists, in the metaphysical sense, appears to depend on prior knowledge of what exists, in the ordinary sense.

Since the agent cannot dispense with the ordinary beliefs about what exists, only sustaining the joint-carving ones, contradictions start to appear in her belief set, at the level of coarse-grained content. This creates the need to understand the connections between the apparently contradictory facts represented by her beliefs. That inquiry cannot be conducted in Ontologese; the agent’s language must include the ordinary quantifier, so that she can inquire into the connections between the two sorts of quantificational facts. So, the best language for inquiring into what exists, for an agent who already has non-joint-carving true beliefs about the matter, is not Ontologese.