KRITERION JOURNAL OF PHILOSOPHY. Volume 29, Issue

Transcription

1 KRITERION JOURNAL OF PHILOSOPHY Volume 29, Issue Johannes Korbmacher: Yet Another Puzzle of Ground Jack Yip: Truthmaking as an Account of How Grounding Facts Hold Jonathan Egeland Harouny: Metaontological Deflationism in the Aftermath of the Quine-Carnap Debate Nicola Claudio Salvatore: Wittgensteinian Epistemology and Cartesian Skepticism Philipp Kanschik: Why Sufficientarianism is not Indifferent to Taxation

2 EDITORIAL KRITERION Journal of Philosophy is a forum for contributions in any field of analytic philosophy. We welcome submissions of previously unpublished papers, not under consideration for publication anywhere else. Submissions are reviewed in double-blind peer review mode. Contributions should meet the following conditions: (1) The content must be philosophical. (2) The language must be intelligible to a broader readership. (3) The contribution must contain a traceable argumentation. The length should be between 4000 and 8000 words. Only contributions in English (preferred) and German are accepted. IMPRESSUM Editors-in-Chief: Christian J. Feldbacher, Alexander Gebharter Editorial Board: Albert J. J. Anglberger, Laurenz Hudetz, Christine Schurz, Christian Wallmann Address: Franziskanergasse 1, 5020 Salzburg, Austria. editor@kriterion-journal-of-philosophy.org Web: Indexing: KRITERION Journal of Philosophy is indexed and abstracted by The Philosopher s Index and EBSCOhost Humanities Source. Information about the journal s ranking is available at SJR. The journal was also approved of satisfying the ERIH (European Reference Index for the Humanities) criteria: ERIH PLUS. Copyright: The copyright remains with the authors. ISSN:

3 Yet Another Puzzle of Ground Johannes Korbmacher Abstract We show that any predicational theory of partial ground that extends a standard theory of syntax and that proves some commonly accepted principles for partial ground is inconsistent. We suggest a way to obtain a consistent predicational theory of ground. Keywords: grounding, standard theory of ground Fine defines ground as the relation of one truth holding in virtue of others [7, p. 1]. 1 Given this definition, it is natural to think that we should formulate axiomatic first-order theories of ground, which formalize ground by means of a relational ground predicate of true sentences. Call such theories predicational theories of ground. Predicational theories of ground contrast with operational theories of ground, which formalize ground by means of a sentential ground operator [3, p , 5, p ]. So far, most theories of ground in the literature are operational theories of ground. But there are at least three theoretical reasons for developing predicational theories of ground: 1. Quantification: Predicational theories of ground have greater expressive strength than operational theories of ground. In particular, using a ground predicate, we can formalize ground-theoretic principles involving quantification over truths in a natural way. Take, for example, the intuitively plausible claim that every truth is either fundamental or grounded in some other truths. We can straightforwardly formalize this claim using a ground predicate and first-order quantification over truths, but using a ground operator this is impossible. Without the use of non-classical devices, such This paper won the SOPhiA 2015 Best Paper Award at the Salzburg Conference for Young Analytic Philosophy, September 2 4, Kriterion Journal of Philosophy, 2015, 29(2): c 2015 The author

4 2 KRITERION Journal of Philosophy, 2015, 29(2): 1 10 as propositional quantification, it is impossible to formalize the nested universal and existential quantification over truths in the principle. Using a ground predicate, in contrast, we can formalize the principle comfortably in the purview of classical first-order logic. 2. Truth and Modality: Predicational theories of ground allow us to study ground in the same context as truth and modality. It is generally accepted that truth should be treated as a predicate of sentences, and it has recently been suggested to extend this approach to modality as well [10, 13, 8]. There is an obvious connection between ground and truth, since ground is a relation among truths. But there is also a close connection between ground and modality, since ground is usually assumed to imply necessary consequence: if a truth holds in virtue of some other truths, then the former truth should be a necessary consequence of the latter truths [5, p ]. Both of these connections are most naturally studied using predicational theories of ground: by combining predicational theories of ground with predicational theories of truth and modality. 3. Models: Predicational theories of ground allow us to discover and to study models of ground using classic model-theoretic methods. It is currently an open problem to provide a semantics for the impure logic of ground developed by Fine [5, p ]. This logic is formulated using a ground operator, but once we translate it into a predicational theory of ground and show its consistency, we can rely on model-theoretic theorems to establish the existence of first-order models. Once we know that such models exist, we can study them using methods of model theory. This should provide us with new insights into the semantics of the impure logic of ground. But predicational theories of ground face a paradox of self-reference, similar to the well-known paradoxes of self-reference that arise in predicational theories of truth and modality. In this paper, I shall prove this point for predicational theories of partial ground in particular. This is the relation of one truth holding partially in virtue of another truth the relation of one truth helping to ground another truth [5, p. 50]. I show that any predicational theory of partial ground that extends a standard theory of syntax and that proves some commonly accepted principles for partial ground is inconsistent. Fine [6] and Krämer [11] present puzzles about the irreflexivity of partial ground: the principle that no truth partly grounds itself. They show that certain intuitively

5 Johannes Korbmacher: Yet Another Puzzle of Ground 3 plausible principles of logic and metaphysics lead to counterexamples to the irreflexivity of ground. I add yet another puzzle of ground to the mix. The new puzzle does not mention the irreflexivity of ground or metaphysical principles unrelated to ground, thus it is genuinely different from the previously known paradoxes. To formulate a predicational theory of partial ground, we first need a theory of syntax that allows us to talk about sentences. 2 It is well-known that we can develop such a theory in any sufficiently strong background theory, like Robinson arithmetic for example. For the present purpose, however, our concrete choice of background theory does not matter. All that matters is that our background theory Θ satisfies the following three minimal syntax conditions: 3 The first condition is that Θ proves that we have a unique name ϕ for every sentence ϕ in the sense that for all sentences ϕ and ψ, Θ ϕ = ψ only if ϕ = ψ. The second condition is that Θ proves that we have a function symbol that represents the syntactic operation of disjunction in the sense that for all sentences ϕ and ψ, Θ ϕ ψ = ϕ ψ. And the third condition is that Θ proves the diagonal lemma. Informally, this lemma states that for every condition on sentences there is a sentence that is provably equivalent to the condition holding of itself. More precisely, if ϕ(x) is a formula with exactly one free variable, then there exists a sentence δ such that Θ δ ϕ( δ ). Note that any standard background theory of syntax, such as Robinson arithmetic, satisfies all three of our minimal syntax conditions. Next, we need a way of representing partial ground. For this purpose, we use the relational predicate x y. For sentences ϕ and ψ, we informally read the atomic formula ϕ ψ as saying that the truth of ϕ partially grounds the truth of ψ. For a negated atomic formula of the form ( ϕ ψ ) we also write ϕ ψ, which we correspondingly read as saying that the truth of ϕ does not even partially ground the truth of ψ. Philosophers have laid down various principles for partial ground [cf. 17, 5, 7], but it is already sufficient for a predicational theory of partial ground to be inconsistent that it proves two widely accepted principles. Let Θ now be a predicational theory of partial ground that satisfies the minimal syntax conditions. The first of our two principles follows directly from partial ground being a relation of true sentences: If the truth of one sentence partially grounds the truth of another, then both sentences should be true. This principle is known as the factivity of ground and is generally accepted in the literature on ground [6, p. 100, 1, 3]. 4 We get the condition on Θ that for all sentences ϕ and ψ:

6 4 KRITERION Journal of Philosophy, 2015, 29(2): 1 10 (Fact L ): (Fact R ): Θ ϕ ψ ϕ Θ ϕ ψ ψ The second principle concerns the interaction of partial ground and disjunction: Given that partial ground is the relation of one truth holding partially in virtue of another, if a disjunction is true, then its truth should be partially grounded in each of its true disjuncts. Also this principle is generally accepted in the literature on ground [6, p. 101, 17, p. 117]. From this, we get the condition on Θ that for all sentences ϕ and ψ: ( 1 ): Θ ϕ ϕ ϕ ψ ( 2 ): Θ ψ ψ ϕ ψ The minimal syntax conditions and the conditions concerning partial ground all may seem fairly uncontroversial when viewed individually. So it may be somewhat surprising to learn that there can be no consistent predicational theory of partial ground that satisfies all of them: Theorem (Inconsistency Theorem). Any theory Θ that satisfies the minimal syntax conditions, (Fact L/R ), and ( 1/2 ) is inconsistent. Proof. Let ϕ(x) be the formula x x x. By the diagonal lemma, there is a sentence δ such that Θ δ δ δ δ. Intuitively, this is a sentence which says of itself that it does not partially ground its own disjunction. By the second minimality condition, we have that Θ δ δ = δ δ. From this and Θ δ δ δ δ, we get that Θ δ δ δ δ by the substitutivity of identicals. This splits up into the following two conditions: (a) Θ δ δ δ δ (b) Θ δ δ δ δ We get finally the following argument: 1. Θ ( δ δ δ δ) ((δ δ δ δ ) δ δ δ ) (Tautology ) 2. Θ δ δ δ δ (Fact L) 3. Θ (δ δ δ δ ) δ δ δ (1, 2: MP) 4. Θ δ δ δ δ (a) 5. Θ δ δ δ (3, 4: MP) 6. Θ δ δ δ δ (b) 7. Θ δ (5, 6: MP)

7 Johannes Korbmacher: Yet Another Puzzle of Ground 5 8. Θ δ δ δ δ ( 1) 9. Θ δ δ δ (7,8: MP) 10. Θ (5,9: ) ( ) : Note that every sentence of the form (ϕ ψ) ((ψ ϕ) ϕ) is a classical tautology and that theories prove all classical tautologies. The inconsistency theorem is very similar to Tarski s theorem about predicational theories of truth [18] and Montague s theorem about predicational theories of modality [15] in that it is, essentially, a paradox of self-reference. From a technical perspective, it should in fact not be surprising that we get such a theorem after all: Combining self-reference via the diagonal lemma with principles like (Fact L/R ) that allow us to push a sentence outside the scope of a predicate and principles like ( 1/2 ) that allow us to push a sentence into the scope of a predicate is recipe for disaster. 5 But from a philosophical perspective, there is a lesson to be learned: We already know that we cannot understand ground simply in terms of truth and modality [5], but ground behaves syntactically too much like a combination of truth and modality to escape inconsistency when paired with self-reference. Three natural ways in which we could try to block the inconsistency theorem suggest themselves: First, we could try to rule out selfreferential sentences of ground like the one used in the proof of the inconsistency theorem. Second, we could try to restrict the principles of partial ground used in the proof of the inconsistency theorem. And third, we could try to formulate a non-standard logic of ground that does not sanction the logical principles used in the proof of the inconsistency theorem. The analogy between the inconsistency theorem and the theorems of Tarski and Montague suggests a terminology for these approaches. Analogously to predicational theories of truth [9] and predicational theories of modality [8], we get: typed theories of partial ground, which avoid paradox by putting type-restrictions on the relation of partial ground, effectively ruling out self-referential sentences like the one in the proof; untyped theories of partial ground, which avoid paradox by restricting the principles of partial ground; and finally non-classical theories of partial ground, which avoid paradox (or: triviality) by abandoning classical logic in favor of alternative logics. Untyped theories of partial ground are particularly appealing, because considerations of ground are already part of intuitively appealing approach to untyped theories of truth. On an influential view about

8 6 KRITERION Journal of Philosophy, 2015, 29(2): 1 10 predicational theories of truth, self-referential sentences are ungrounded and this is the reason some self-referential sentences lead to inconsistency [12, 14]. This leads to the idea that we should restrict the principles of truth to their grounded instances. 6 Carrying this idea from theories of truth over to predicational theories of ground, we arrive at the condition that the principles of partial ground apply if and only if the truths involved are themselves grounded. There is a straightforward way of formulating the desired restriction on the principles of partial ground already in the language of partial ground. We can express that a sentence ϕ is grounded by the formula x(x ϕ ) and we can express that a sentence ϕ is ungrounded by the formula x(x ϕ ). The desired restriction on ( 1/2 ) then amounts to saying that for all predicational theories of ground Θ and for all sentences ϕ and ψ: ( 1): Θ x(x ϕ ) ϕ ϕ ψ ( 2): Θ x(x ψ ) ψ ϕ ψ Every predicational theory of partial ground that satisfies the minimal syntax conditions, (Fact L/R ), and the new conditions ( 1/2 ), proves that the paradoxical sentence in the proof of the theorem is ungrounded: Observation. Let Θ be a predicational theory of ground that satisfies the minimal syntax conditions, (Fact L/R ), and ( 1/2 ). By the diagonal lemma and the same reasoning as in the proof of the theorem, we get a sentence δ such that: Θ δ δ δ δ. But we can show that: Θ x(x δ ). Proof. By applying ( 1) to δ, we get that: Θ x(x δ ) δ δ δ We only need the left-to-right direction of this biconditional for our proof, which we can obtain via -Elimination: Θ x(x δ ) δ δ δ Starting from there, we get the following argument: 1. Θ x(x δ ) δ δ δ 2. Θ δ δ δ δ (Fact L)

9 Johannes Korbmacher: Yet Another Puzzle of Ground 7 3. Θ x(x δ ) δ (1,2: MP) 4. Θ δ δ δ δ (Diagonal Lemma) 5. Θ x(x δ ) δ δ δ (3, 4: -Elim) 6. Θ x(x δ ) (1, 5: -Intro) 7. Θ x(x δ ) (6: -Intro) This result should make us optimistic about the prospects for a untyped theory of partial ground. Moreover, we could add such a untyped theory of partial ground on top of untyped predicational theories of truth and modality. I conjecture that an interesting, consistent, untyped predicational theory of ground can be developed in this way. Acknowledgements I would like to thank Albert J. J. Anglberger, Hannes Leitgeb, Thomas Schindler, and Ole Thomassen Hjortland for helpful comments and suggestions. Notes 1 For (opinionated) introductions to ground, see [4, 5]. For an overview of the recent literature on ground, see [2, 19, 16]. 2 A sentence is a formula without any free variables. 3 A theory is a set of formulas that is closed under derivability: a set of formulas Θ is a theory iff (if and only if) for all formulas ϕ, if Θ ϕ, then ϕ Θ. 4 There are notions of ground in the literature that violate the factivity of ground [cf. 5, p ]. According to such non-factive notions, ground is a relation on sentences regardless of their truth value. Although non-factive notions of ground make for an interesting theoretical possibility, in this paper we shall deal only with the standard factive notion of ground, which satisfies the factivity of ground. 5 It should be clear at this point that not much depends on the concrete condition ( 1/2 ) the paradox is not a paradox of disjunction. All that matters is that our predicational theory of partial ground proves a principle to the effect that any true sentence partially grounds some other sentence. We could give the following variant of the inconsistency theorem: If Θ satisfies (Fact L/R ) and either Θ ϕ x( ϕ x) or Θ ϕ x(x ϕ ), then Θ is inconsistent. I leave the details of the proof to the interested reader. 6 The concept of ground used in the context of theories of truth is not exactly the same as the concept of ground discussed in this paper. For example, the notion of dependence defined by Leitgeb [14] is reflexive, whereas (partial) ground is standardly taken to be irreflexive. The point here is that there is a striking

10 8 KRITERION Journal of Philosophy, 2015, 29(2): 1 10 analogy between the two concepts and that ideas that work for the one may very well work for the other. Johannes Korbmacher MCMP, LMU Munich Geschwister-Scholl-Platz Munich, Germany <jkorbmacher@gmail.com> < References [1] Ricki Bliss and Kelly Trogdon. Metaphysical Grounding. In: The Stanford Encyclopedia of Philosophy. Ed. by Edward N. Zalta. Winter [2] Michael Clark and David Liggins. Recent Work on Grounding. In: Analysis 72.4 (2012), pp [3] Fabrice Correia. Grounding and Truth-Functions. In: Logique et Analyse (2010), pp [4] Fabrice Correia and Benjamin Schnieder. Grounding. An Opiniated Introduction. In: Metaphysical Grounding: Understanding the Structure of Reality. Ed. by Fabrice Correia and Benjamin Schnieder. Cambridge, UK: Cambridge University Press, 2012, pp [5] Kit Fine. Guide to Ground. In: Metaphysical Grounding: Understanding the Structure of Reality. Ed. by Fabrice Correia and Benjamin Schnieder. Cambridge, UK: Cambridge University Press, 2012, pp [6] Kit Fine. Some Puzzles of Ground. In: Notre Dame Journal of Formal Logic 51.1 (2010), pp [7] Kit Fine. The Pure Logic of Ground. In: Review of Symbolic Logic 5.1 (2012), pp

11 Johannes Korbmacher: Yet Another Puzzle of Ground 9 [8] V. Halbach and P. Welch. Necessities and Necessary Truths: A Prolegomenon to the Use of Modal Logic in the Analysis of Intensional Notions. In: Mind (2009), pp [9] Volker Halbach. Axiomatic Theories of Truth. Cambridge, UK: Cambridge University Press, [10] Volker Halbach, Hannes Leitgeb, and Philip Welch. Possible- Worlds Semantics for Modal Notions Conceived as Predicates. In: Journal of Philosophical Logic 32.2 (2003), pp [11] Stephan Krämer. A Simpler Puzzle of Ground. In: Thought: A Journal of Philosophy 2.2 (2013), pp [12] Saul A. Kripke. Outline of a Theory of Truth. In: Journal of Philosophy (1975), pp [13] Hannes Leitgeb. Towards a Logic Of Type-Free Modality and Truth. In: Logic Colloquium 2005: Proceedings of the Annual European Summer Meeting of the Association for Symbolic Logic, Held in Athens, Greece, July 28-August 3, Ed. by Costas Dimitracopoulos. Cambridge University Press, 2008, pp [14] Hannes Leitgeb. What Truth Depends On. In: Journal of Philosophical Logic 34.2 (2005), pp [15] Richard Montague. Syntactical Treatment of Modality, with Corollaries on Reflection Principles and Finite Axiomatizability. In: Acta Philosophica Fennica 16 (1963), pp [16] Michael J. Raven. Ground. In: Philosophy Compass 10.5 (2015), pp [17] Gideon Rosen. Metaphysical Dependence: Grounding and Reduction. In: Modality: Metaphysics, Logic, and Epistemology. Ed. by Bob Hale and Aviv Hoffmann. Oxford University Press, 2010, pp [18] Alfred Tarski. Der Wahrheitsbegriff in den Formalisierten Sprachen. In: Studia Philosophica 1 (1936), pp [19] Kelly Trogdon. An Introduction to Grounding. In: Varieties of Dependence. Ontological Dependence, Grounding, Supervenience, Repsonse-Dependence. Ed. by Miguel Holtje, Benjamin Schnieder, and Alex Steinberg. München: Philosophia Verlag, 2013, pp

12

13 Truthmaking as an Account of How Grounding Facts Hold Jack Yip Abstract Grounding, as a way to articulate ontological dependence, faces the problem of what grounds grounding facts themselves (such as the fact that the singleton of Socrates is grounded in Socrates). This problem stems from the need to account for the holding of grounding facts, which generates the hierarchical structure of ontological dependence. Within the grounding framework, grounding facts are either ungrounded or grounded. I will first argue that neither option can provide us with a satisfactory account. The main reason is that non-fundamental entities have to be counted as fundamental or involved in the essences of fundamental entities in order for either of the two options to work the non-fundamental is being smuggled into the fundamental. My suggestion is to appeal to the notion of truthmaking and tackle the problem about the holding of grounding facts outside the grounding framework instead of asking what grounds grounding facts, I ask what makes grounding claims true. Truthmaking is a prima facie relation holding between the representational and the non-representational such that the latter makes the former true. With the principle if p is true, then it is a fact that p, we can account for the holding of grounding facts in a derivative sense. As a proposition contains the information about its truthmaker, the nature of grounding claims will tell us how grounding facts hold. I accept a realm of concepts which make up propositions (which might be needed already if there are propositions and propositions are compositional). These concepts will act as part of the truthmaker for grounding claims (in addition to the non-conceptual fundamental entities) the concept of the ground must figure in the concept of the grounded. For a concept to figure in another, it is to be involved in the constitutive essence of the latter (analogous to Kit Fine s idea that the ground of a grounded entity figures in the essence of the grounded entity). This account will not smuggle anything non-fundamental into the fundamental realm. The implication is that ontological dependence stems from our different kinds of conceptualisations (perhaps of the same stuff, as in the concepts of water and H 2O), which justifies metaphysicians armchair method. Keywords: truthmaking, grounding, essence of ground. entities Kriterion Journal of Philosophy, 2015, 29(2): c 2015 The author

14 12 KRITERION Journal of Philosophy, 2015, 29(2): A sketch of grounding Grounding is a primitive notion employed by many contemporary metaphysicians to articulate ontological dependence. It is devised in light of the realisation that intensional or modal notions are not fine-grained enough to do the job. The classic example is given by Kit Fine ([4]; [5]): Socrates and the singleton of Socrates. While both of them supervene upon each other or necessitate the existence of each other, it is the singleton of Socrates that is ontologically dependent upon Socrates but not vice versa 1. So, modal notions fail to capture the asymmetry involved. Grounding is a hyperintensional notion philosophers come up with to replace such modal notions in order to formulate claims about ontological dependence. The ontological dependence of the singleton of Socrates upon Socrates lies in the fact that the singleton of Socrates is grounded in Socrates. All such grounding facts together generates a hierarchical structure of ontological dependence Kit Fine writes, one might first ground the normative in the natural, for example, then the natural in the physical, and then the physical in the micro-physical, thereby establishing that the normative was grounded in the micro-physical. ([7], p.44) What figures at the bottom of the structure is usually taken to be ungrounded, fundamental entities ([19]). There are largely two different formulations of grounding facts or grounding claims. The first takes the non-truth-functional sentential operator x because y with x and y to be substituted by sentences as the canonical formulation 2 such that what the sentence in the place of x expresses is grounded in what the sentence in the place of y expresses. For example, the singleton of Socrates is what it is because Socrates is what it is. This formulation is endorsed by Kit Fine ([7]; [6]). It has the virtue of metaphysical neutrality. First, this formulation does not commit one to a grounding relation. Second, it does not imply the existence of the grounded and the ground. Third, it does not tell us whether the grounded is to be reduced to the ground even though the grounding claims would be a necessary condition for the corresponding reductions. The second formulation takes the relational predicate y grounds x as canonical. Gideon Rosen ([18]) restrains the relata to facts, while Jonathan Schaffer ([19]) allows different kinds of entities to figure in them. This formulation would prima facie commit one to a grounding relation and also to the existence of its relata. But reduction becomes a problem as the grounding relation is irreflexive this problem leads to

15 Jack Yip: Truthmaking and Grounding 13 the more complicated formulations of grounding as a quaternary relation ([11]; [20]). No matter which formulation one adopts, there is a problem that confronts all grounding theorists what grounds the grounding facts or grounding claims 3 themselves? We know that the singleton of Socrates is grounded in Socrates. But what makes it so? It should be a matter of fact that something is grounded in something else. But what grounds do such grounding facts have? How can they hold if they are not grounded? How can we tell whether a purported grounding fact holds or not? 2 The problem of what grounds grounding facts As noted in the previous section, all the grounding facts together generate a hierarchical structure of ontological dependence with fundamental entities at the bottom. The holding of this structure is predicated upon the holding of those grounding facts. But how can they hold? It is intuitively true that the fact that London is a city is grounded in certain facts concerning the complex physical structure of the area designated by London. 4 But why? Why can t it be the reverse? Why can t it fail to hold? This grounding fact and others are essential to the standing of the hierarchy of ontological dependence. So, an account of the holding of such grounding facts is necessary to the account of ontological dependence. There are two options within the grounding framework: Grounding facts are either grounded or ungrounded 5. I will consider the second option first. Let s use the following names for our exposition: L = the fact that London is a city P = the whole of the facts concerning the complex physical structure of the area designated by London (which we assume to be fundamental) G = the fact that L is grounded in P. According to the second option, grounding facts are ungrounded. So, in our example, G is ungrounded. The problem is that under any reasonable conception of fundamentality, G would then be fundamental. 6 But, G contains a non-fundamental component, i.e. L, provided that cities are not fundamental. Given that something fundamental cannot contain any non-fundamental component ([6]; [21], ch ), G is not fundamental. Therefore, G should not be ungrounded.

16 14 KRITERION Journal of Philosophy, 2015, 29(2): This objection presupposes that a grounding fact is composed of a grounding relation and its relata in a certain way (perhaps as a nonmereological composition ([1])). But it might be argued that we can take grounding facts as atomic such that L and P would not be components of G in the above example. So, even if grounding facts are ungrounded, they would not contain any non-fundamental component, and thus they would not be counted as non-fundamental. 7 This atomic conception of grounding facts resembles what David Lewis calls the magical conception of structural universals ([14]). Given the factivity of grounding, a grounding fact necessitates the holding of the ground and the grounded. In my original presupposition, it s easy to see how this necessitation works, as a grounding fact is composed of a grounding relation and its relata. But with grounding facts being atomic, they do not involve the grounded and the ground as their components. Thus the grounding facts, the grounded, and the ground seem to be wholly distinct entities. It s hard to see how such necessitations can hold between them it violates the Humean principle that there is no necessary connection between wholly distinct entities. Such necessitations can only hold as a stipulated brute modal fact a mysterious one. Moreover, there is the question about how a certain grounding fact can be identified as G. Calling it G does not make it so. There seems to be no way to distinguish one particular grounding fact G from other grounding facts. Perhaps we can postulate some further fact about the necessary connection among G, P and L to account for the necessitation and the identification of G. But then the question about the holding of this further fact will have to be raised again if it s atomic, it falls victim to the arguments in this paragraph; if it s composed of G, P, and L in a certain way, then it will contain some non-fundamental component, so it can t be ungrounded. 8 Furthermore, even if ungrounded atomic grounding facts somehow succeed in doing the work they are supposed to do, we can still point out another involvement of non-fundamental entities in the fundamental realm the essence of an ungrounded grounding fact (i.e. what it is for a grounding fact to be that grounding fact) must involve entities that are not fundamental. (My conception of essence follows from Kit Fine s see n.12.) So, there are at least two ways to cash out the idea that an ungrounded grounding fact unduly involves non-fundamental entities. But why can t something fundamental contain or involve anything non-fundamental? I have no room to argue for it, so I will only put it metaphorically: God does not have to create cities in order for the

17 Jack Yip: Truthmaking and Grounding 15 world to have cities God only creates what is fundamental and everything non-fundamental follows. Or, if we had a language that completely carves at the joints of reality, expressions referring to facts about cities would not figure in that language ([21]). Without any principled reason to suggest otherwise, fundamental grounding facts should not be exempted from this constraint. The above strategy which takes grounding facts as ungrounded is mentioned by Schaffer ([19], p.373, n.32) in passing. He says that grounding stands outside the priority ordering [i.e. the aforementioned hierarchy] altogether, imposing structure upon it. But according the Schaffer s own definition of fundamentality for any x, x is fundamental = df nothing grounds x ([19], p.373) such grounding facts must be fundamental. Supposing that G is ungrounded, it is fundamental by definition. But G is not fundamental. So G should not be ungrounded. As G is just an arbitrarily picked grounding fact, the same should hold for other grounding facts as well (unless there were some grounding relations holding only between fundamental entities 9 ). This objection hinges on a conception of fundamentality which defines it in terms of grounding. But there is an alternative proposal by Kit Fine, which takes the notion of fundamentality to be primitive. The motivation behind this approach is that fundamentality is an absolute notion, and thus it should not need a relational underpinning. Fine thinks that it is simply the conception of Reality as it is in itself. ([6], p.25) This notion of fundamentality is connected to other notions. Fine defines what is basic as what is ungrounded. But what is basic need not be fundamental. For example, that abortion is wrong can be taken as basic, but there might be no fact in reality which dictates that abortion is wrong. So, the wrongness of abortion might not be factual, while it is still ungrounded. A sufficient condition for something to be fundamental is for it to be basic and factual. According to Fine, for something to be factual is for it to be factual itself or for it to be grounded in something factual. Given that G is ungrounded, it must be factual itself in order for it to be factual. Can we avoid the fundamentality of G by taking it as basic but nonfactual? If G is nonfactual, then there is nothing in reality that corresponds to G. More generally, grounding claims would then not be representational of reality, and they would just be a projection of our conception of reality onto reality, or, to put it in another way, they would just reflect our practice of talking about ontological dependences, while there are no such dependences in reality. Using Ted Sider s terminology ([21]), grounding facts would then not be a metaphysically substantive

18 16 KRITERION Journal of Philosophy, 2015, 29(2): matter (which is about reality in itself), but instead merely a conceptually substantive matter (which is about our conceptual scheme). As grounding is intended to carve out the hierarchical structure of ontological dependence in reality ([19]), grounding should be metaphysically substantive. Therefore, grounding facts should be factual rather than nonfactual. Given that grounding facts are factual and ungrounded (i.e. basic), they are fundamental. So, this alternative conception of fundamentality cannot avoid the fundamentality of G, given its ungroundedness. As a result, G should not be ungrounded. What if we take grounding facts to be grounded? This approach does not commit us to the fundamentality of grounding facts, so it avoids the objection above. But the next question is: in what are grounding facts grounded? There are at least four possible candidates: (1) A grounding fact is grounded in the ground involved in itself. (2) A grounding fact is grounded in the essence of the grounded entity involved. (3) Some grounding facts are fundamental and they ground other nonfundamental grounding facts. (4) A grounding fact is grounded in both the grounded and the ground, i.e. it concerns an internal relation which holds in virtue of the holding of its relata. I will consider each of them one by one. 10 (1) Louis derosset ([3]) takes the ground involved in a grounding fact as the ground of that grounding fact itself. Using our example, G would be grounded in P. But here we have another grounding fact apart from G, i.e. the fact that G is grounded in P ( G+ ). Following the principle that the ground in a grounding fact grounds the grounding fact itself, G+ is also grounded in P. Now we have another grounding fact, i.e. the fact that G+ is grounded in P ( G++ ). Following the same principle, G++ is grounded in P. This chain of grounding facts will go on ad infinitum. All of them will be grounded in P. 11 A question arises: what is it about P that makes it the ground of all these grounding facts? For all these grounding facts to hold, there must be something in the fundamental reality that grounds them. Under the supposition that they are grounded in P, there must be something about P that makes it the ground of these grounding facts. It is reasonable to think that it is in the essence or nature of P that it grounds these grounding facts. If so, it would be in the essence of P that P grounds G.

19 Jack Yip: Truthmaking and Grounding 17 As G contains a non-fundamental component L, the essence of P would involve something non-fundamental. 12 Given that the non-fundamental cannot be involved in the fundamental, the essence of P cannot involve facts about cities. So, it fails to ground G and other grounding facts in the chain. It might be argued that P can ground G and other grounding facts without any involvement of these grounding facts in its essence. So, while P grounds these grounding facts, P could hold without grounding these grounding facts. But grounding (at least in the metaphysical sense under consideration) implies metaphysical necessity ([7], sect.1.1). If P can exist without grounding these grounding facts, then P is not the ground for them. The reason is that there is some possible world where P holds without the holding of L, G, and other grounding facts in the chain. To consider an example, water is grounded in H 2 O. But it would be wrong to assert this grounding claim, if H 2 O can exist without the existence of water. derosset ([3]) argues that we should distinguish a grounding story (what he calls an explanatory story ) from a ground with regard to the same grounding fact. He provides an analogy: the fact that it is chilly or windy is grounded in the fact that it is chilly. In order to make this grounding claim, some ancillary material is needed, such as something that concerns the nature of disjunctive facts, which might say that a disjunctive fact holds iff at least one of its disjuncts holds. But this ancillary material is not a part of the ground for the disjunctive fact only its disjuncts can be the ground. In his analogous example, the nature of disjunctive facts is not part of the ground for the fact that it is chilly or windy. The ground for it is just the fact that it is chilly. The nature of disjunctive facts figures only in the grounding story. Analogously, the connection between cities and certain complex physical structures need not figure in the ground for those cities. It can just be a part of the grounding story. So, in our example, non-fundamental elements like facts about cities need not be involved in the essence of P. For the things that figure in the grounding story to hold in reality, they must be either fundamental or grounded in the fundamental. For otherwise, the grounding facts would not hold in reality, given their partial dependence on the grounding story. In the case of the nature of disjunctive facts, we can take it as a general fundamental metaphysical law, or we can take it as the essence of everything fundamental at all everything is such that for any disjunctive fact, it is grounded in at least one of its disjuncts.

20 18 KRITERION Journal of Philosophy, 2015, 29(2): But this treatment is not applicable to the grounding story needed in G. To have something akin to the notion of disjunction in the fundamental is not implausible for example, Ted Sider might want to have it as part of the structure of reality, out of the consideration that the notion of disjunction might have to figure in the robustly best theory of the world ([21]). But to have facts about cities in the fundamental is implausible. So, in order for the grounding story for G to hold, the connection between cities and certain complex physical structures must be grounded in something fundamental rather than being itself fundamental. But what can ground it except for P? Then again, we would have the involvement of facts about cities in the essence of P, which is implausible. Therefore, P cannot be the ground for G. More generally, grounding facts cannot be grounded in the grounds involved in them. (2) Kit Fine ([7], sect.1.11) proposes to have the essence of the grounded involved in a grounding fact to be the ground for that grounding fact itself. Using our example, G would be grounded in the essence of L. But we have another grounding fact apart from G, i.e. the fact that G is grounded in the essence of L ( G* ). Following the principle proposed by Fine, G* is grounded in the essence of G. Now we have another grounding fact, i.e. the fact that G* is grounded in the essence of G ( G** ). Following the same principle, G** is grounded in the essence of G*. Apparently, this chain will go on ad infinitum. 13 Fine motivates this strategy on the basis of his conception of the methodology of metaphysics. He thinks that part of the investigation into ground is the investigation into the essences of things. And, it is by discovering what the essences of the things to be grounded are that we find out what their grounds should be. This approach raises the question about what grounds the essences of the grounded involved in the chain of grounding facts. If the grounding facts are to hold in reality, the essences that ground them must be either fundamental or grounded in something fundamental. Given that the essences of the non-fundamental are not fundamental, these essences should be grounded in the fundamental. Consider the essence of L which grounds G. As L is grounded in P, it is natural to think that the essence of L is also grounded in P. But then it seems that P has no way to do this except for having the involvement of the essence of L in its own essence. This would amount to smuggling the non-fundamental into the fundamental. There seems to be no other plausible ground for the essence of L (let alone the essences of G, G*, G**,..., ad infinitum) which does not involve this kind of smuggling. So, Fine s approach is

21 Jack Yip: Truthmaking and Grounding 19 implausible. (3) The third possible approach takes some grounding facts as fundamental and it is these fundamental grounding facts which ground other grounding facts. There are at least two problems about this approach. First, it is unclear how fundamental grounding facts can ground nonfundamental ones. The fundamental grounding facts are supposed to hold between fundamental entities. For example, in an ontology with Aristotelian universals and thick particulars as fundamental entities, they should stand in fundamental grounding relations to one another. Suppose that physics tells us what counts as fundamental properties and particulars. It is difficult to see how such fundamental grounding facts can ground the grounding facts holding between cities and fundamental universals plus particulars. For them to do so, they should involve non-fundamental grounding facts or entities in their essences. So, this leads us to the second problem for this approach to work, one has to smuggle the non-fundamental into the fundamental. (4) The last possible approach is to take grounding as an internal relation. An internal relation holds in virtue of the existence of its relata. The classic example is resemblance: given two objects, any resemblance relations between them hold merely in virtue of the existence of the two objects. So, these resemblance relations are not anything ontologically over and above the two objects in question. In the case of grounding facts, we can say that grounding facts are grounded in the holding of both the grounded and the ground involved. In our example, G is grounded in the holding of both L and P. The problem with this approach is that it gets the ontological priority wrong we won t have the grounded entity unless the grounding fact holds. The grounded holds only in virtue of its being grounded in the ground, i.e. the holding of the corresponding grounding fact, and the holding of the ground. We cannot say that the grounding fact holds in virtue of the grounded and the ground involved. For example, L holds in virtue of the holding of G and the holding of P. We cannot say that G holds in virtue of the holding of L and P. That would result in the wrong order. So, this last approach doesn t work. The above considerations lead us to an unfortunate result: under the framework of grounding, we cannot account for how grounding facts hold, no matter whether they are taken to be grounded or not. But the holding of the grounding facts is what gives us the hierarchical structure of ontological dependence. If their holding is in question, the structure will be in question as well. The grounding talk is used to replace modal

22 20 KRITERION Journal of Philosophy, 2015, 29(2): notions to articulate theses of ontological dependence. But if we cannot explain why and how grounding facts hold, the grounding talk will only let in a novel but suspicious way of speaking. For example, we want to say that the singleton of Socrates is grounded in Socrates this is how we articulate the ontological dependence between them. But if when asked about how this grounding fact holds, we have no satisfactory answer, then it would seem that we have no way to assess the truth of different grounding claims, while we have a clear grasp of the ontological dependence of the singleton of Socrates upon Socrates. This points to a discrepancy between our preferred articulation of a thesis and the wanted thesis itself, which is going to undermine the value of the grounding framework. 3 Truthmaking as a way out From the above considerations, we can see that solving the problem of what grounds grounding facts under a framework of grounding faces insurmountable difficulties. Perhaps a way out can be found by a reconstrual of the problem such that it can be addressed independently of a framework of grounding. Such a reconstrual is not identical to the original problem, but it can address the bigger problem of how grounding facts hold. My suggestion is to make use of the notion of truthmaking. Truthmaking is a prima facie relation that holds between a representational entity (such as a proposition or a truth) and a nonrepresentational entity (such as a fact) such that the representational is made true by the non-representational (or something representational whose representational nature is not essential for it to be the truthmaker) 14. It tells us why a proposition 15 is true, under the supposition that the non-representational determine the truth-values of the representational. For example, the proposition Jack is a man is made true by the fact that Jack is a man (or, if one does not have such a macroscopic fact in one s ontology, then one can have anything that ontologically amounts to this fact as the truthmaker for the proposition). 16 How do we make use of truthmaking to help solve the problem? First, grounding facts can be expressed by propositions. So, G can be expressed by the proposition the fact that London is a city is grounded in the whole of the facts concerning the complex physical structure of the area designated by London. Then with the principle (TF) if the proposition p is true, then it is a fact that p 17

23 Jack Yip: Truthmaking and Grounding 21 we can answer the question about how grounding facts hold by answering the question about what makes grounding claims true. This approach apparently resembles Fine s canonical formulation of grounding sentences of the form p because q. But Fine does not ask the question about what makes such claims true. He simply sticks to the framework of grounding to ask what grounds them, which leads him to the implausible treatment we have considered in the last section (i.e. the grounding facts are to be grounded in the essences of the grounded). But once we adopt the framework of truthmaking, we can address the problem from a different perspective. Instead of asking what grounds a fact, we ask what it is that makes true a proposition. We know that it is Jack that makes Jack exists true, just like we know that the singleton of Socrates is grounded in Socrates. A question similar to what grounds grounding facts arises: what is it that makes it the case that Jack makes Jack exists true? Or more generally, what is it that makes it the case that a truthmaker makes a certain proposition true? The answer lies in the nature of a proposition. To understand a proposition involves the understanding of the circumstances under which it is true. Putting it in metaphysical terms, a proposition should carry enough information within itself about what makes it true and how its truthmaker makes it true. As one way to think of it, a proposition can be taken as a set of possible worlds where that proposition is true what those worlds share in common is the truthmaker of that proposition. (Of course, this conception of propositions collapses all propositions which express necessary truths into one, fails to distinguish distinct propositions 18, and neglects the different ways in virtue of which a proposition can be true 19. Stephen Yablo ([23]) tries to use the subject matter of a proposition as a way to carve up the logical space into divisions which account for the different ways for that proposition to be true or false. He also constructs impossible worlds to deal with the problem created by necessary truths. The notion of proposition here is not predicated upon the success of his project, but I insist that there is a intuitive sense that a proposition should be fine-grained enough to carry enough information about what makes it true and how its truthmaker makes it true.) For example, Jack exists carries enough information to tell us that its truthmaker is Jack, and the existence of Jack is going to make it true. Similarly, grounding claims themselves should carry enough information in their nature to indicate what things make them true and how they do it.

24 22 KRITERION Journal of Philosophy, 2015, 29(2): We face a problem immediately. We know that water exists is true in virtue of the existence of H 2 O. But it isn t the case that the understanding of water exists involves the understanding of the concept of H 2 O. So, the nature of water exists doesn t carry enough information about its truthmaker. The solution to this problem comes from the Kripke-Putnam conception of natural kind terms. Water got its referent fixed in its baptism through ostension, which is done under the supposition that whatever shares the hidden structure of the target of ostension is to be referred to by water ([16]; [12]). We can extend this treatment of the term water to the concept of water. Assuming that the concept of water figures in the proposition water exists, we can understand the proposition water exists without knowing what its truthmaker really is. Knowing what its truthmaker is depends on the a posteriori studies into the hidden structure of water, which is the job of scientists. With Putnam s linguistic division of labour in mind, the definite answer to the question about what makes water exists true is deferred to the scientific experts ([16]). But it is still in the nature of water exists that it is to be made true by the existence of whatever the scientists take to be water, because it is the concept of water that carries the information about the sample-referent of water whose hidden structure is to be discovered by scientists. The people who know that water is H 2 O and the people who don t know that would have the understanding of the same proposition (but not the same understanding). 20 How then do grounding claims carry enough information about their own truthmakers? Take water is grounded in H 2 O as an example. Given that the concept of water refers to a bunch of stuff, if that bunch of stuff turns out to be H 2 O upon some scientific investigations, then the grounding claim is true. Why is it true? My answer is: (1) water is H 2 O, (2) the concept of H 2 O is more natural or fundamental or real 21 than the concept of water; and (3) there is some portion of fundamental reality which matches the nature inherent in the concept of H 2 O or there is some portion of fundamental reality which matches the nature inherent in a certain concept Q such that H 2 O is grounded in Q is true. It is a posteriori investigations that lead us to know (1). But given that water is H 2 O, the concept of water and the concept of H 2 O have the same referent but different senses (the case here is similar to the one

25 Jack Yip: Truthmaking and Grounding 23 between the concept of Phosphorus and the concept of Hesperus) ([8]). What (1) tells us is the identity of reference between the two concepts. But they differ at least in that the concept of H 2 O figures in the essence of the concept of water (see n.22), and this is not the case conversely. But this kind of identity is not sufficient to account for the truth of grounding claims. Generally speaking, the grounded is more fundamental than its ground. (2) tells us exactly this. Moreover, given the factivity of grounding, there should be something in reality which matches the concept of the ground this requirement is fulfilled by (3). I take this matching as primitive (for example, H 2 O matches the concept of H 2 O), similar to the notion of joint-carving in Sider 2011 ([21]). This account of the truth of grounding claims seems to require a rich ontology at the fundamental level. First, we need concepts as abstract representational entities out of which propositions are composed. Second, we need a primitive ordering of different concepts in accordance with their degrees of fundamentality ( the fundamentality ordering ). The concepts that match the the non-conceptual reality in itself are the most fundamental. Third, we need a fundamental ontological base of things in reality. (But we will see later that the fundamentality ordering can be avoided.) What then is the truthmaker for a grounding claim? A truthmaking relation requires the existence of its relata. So, we have propositions that are made up from concepts in our fundamental ontology to stand for the representational entities to be made true. The nature of a proposition is going to tell us what makes it true. According to our reading of grounding claims, it is the fundamentality ordering and the portion of the ontological base matching the concept of the ground which act as the truthmaker for a grounding claim. To apprehend the nature of a proposition need not be an a priori matter, given that propositions are made up of concepts and the nature of certain concepts (like the concept of water) is subject to external factors. For example, even though the ancient people understood the same proposition water exists, they did not have the same understanding as we do, as now we know that water is H 2 O. The concept of water shares the same referent as the concept of H 2 O objectively, but we discovered that to be the case a posteriori. But what if someone asserts the proposition water is grounded in CO 2? The concept of CO 2 is more fundamental than the concept of water (assuming that chemical concepts are more fundamental than lay concepts), and some portion of reality matches it in the sense similar to

26 24 KRITERION Journal of Philosophy, 2015, 29(2): (3). How can we say that it is false? We need a further requirement for the grounding claims to be made true: the concept of the ground should figure in the concept of the grounded (analogous to Fine s idea that the ground of a grounded entity figures in the essence of the grounded entity ([5])). 22 That a concept figures in another can be a matter of a posteriori discovery such as in the case of water is grounded in H 2 O, or a priori conceptualisation such as in the case of the singleton of Socrates is grounded in Socrates. This requirement forces us to include its fulfilment in our truthmaking picture we need facts such as the fact that the concept of H 2 O figures in the concept of water. I would claim the such facts are nothing over and above the concepts involved in it. 23 So, we need to include concepts as the truthmakers for grounding facts. As we already need concepts to make up propositions, so there will not be any addition to our ontology. Another problem arises with the requirement that the concept of the ground should be more fundamental than the concept of the grounded this need not be the case: for example, with a fundamental ontology of Aristotelian universals and thick particulars, fundamental entities might be grounded in one another, and thus there is no difference in the degrees of fundamentality between the concept of the ground and the concept of the grounded. So, our requirement prematurely rules out some plausible cases of grounding. Moreover, it might be wondered whether all nonfundamental concepts can be lined up in a fundamentality ordering. Fortunately, with the requirement that the concept of the ground figures in the concept of the grounded, a fundamentality ordering of concepts is no longer necessary to account for how grounding claims are true. The general principle is that that a concept A figures in another concept B but not vice versa is already a sufficient condition for A s being more fundamental than B. (As far as I know, there is no counterexample to this principle. The concepts that match the non-conceptual reality in itself are the fundamental concepts. They are involved in the essences of less fundamental ones. Assuming that physical concepts are the fundamental ones, they figure in chemical concepts, which in turn figure in biological concepts, and so on. This principle also matches the philosophical method of reduction for example, if we contend that the physical is not more fundamental than the chemical, we usually can argue that either some physical concepts and some chemical concepts figure in one another, or there are some chemical concepts which have

27 Jack Yip: Truthmaking and Grounding 25 no physical concepts figuring in them, i.e. they cannot be analysed in terms of physical concepts.) Given that the concept of H 2 O figures in the concept of water but not vice versa, and a portion of fundamental reality matches the concept of H 2 O in the sense of (3), it is already true that water is grounded in H 2 O. The lesson is that for grounding claims to be true, we need not settle the complete fundamentality ordering of concepts only the part of it which is due to concepts figuring in one another suffices. To account for possible cases of mutual grounding at the fundamental level, the requirement for the concept of the ground to figure in the concept of the grounded should replace the requirement that the former concept is more fundamental than the latter. Consider the case of a fundamental ontology of Aristotelian universals and thick particulars. The concept of a universal must involve the concept of at least one thick particular (an indefinite one) instantiating it in the essence ([5], pp ), and the concept of a particular (taken as a temporal slice) must involve the concepts of all the universals it instantiates in the essence. Given this fundamental ontology, with the fulfilment of the figuring-in requirement by the concepts, they are grounded in one another. Without the need to require the concept of the ground to be more fundamental than that of the grounded, we don t need to have a fundamentality ordering of concepts at the fundamental level. To sum up my approach, I ask what makes a grounding claim true instead of what grounds a grounding fact, which can lead to a solution to the more general problem about how grounding facts hold. A grounding claim is a proposition. The nature of a proposition will tell us what makes it true and how its truthmaker makes it true. In a true grounding claim, (1*) the concept of the ground should figure in the concept of the grounded; and (2*) some portion of fundamental reality matches the nature inherent in the concept of the ground B, or some such portion matches the nature inherent in another concept C such that B is grounded in C is true. The corresponding truthmaker for grounding claims includes: (i) the concept of the ground and the concept of the grounded (ii) the portion of fundamental reality that matches the concept of the ground in the sense indicated in (2*)

28 26 KRITERION Journal of Philosophy, 2015, 29(2): Applied to our example concerning London, the fact that London is a city is grounded in the whole of the facts concerning the complex physical structure of the area designated by London is made true by (a) the concept of the fact that London is a city ( LC ) and the concept or concepts of the whole of the facts concerning the complex physical structure of the area designated by London ( PC ), and (b) the portion of reality that matches PC. Finally, by (TF), it is a fact that L is grounded in P. In comparison with the different approaches under the framework of grounding considered in the last section, my approach has two virtues. First, it does not involve any vicious regress ad infinitum, as the one in Fine s approach. Second, it does not smuggle anything non-fundamental into the fundamental. Of course, now we have concepts at the fundamental realm. But we have to have concepts anyway if we need propositions and propositions are compositional. (Note that having concepts as fundamental entities does not imply that they are not the results of human activities.) Also, it is more reasonable to have something like the concept of city rather than facts about cities at the fundamental level, given the plausible contention that the representational cannot be reduced to the non-representational. Besides, there are well-known independent reasons for having this third realm of representational entities, as noted by Frege ([9]). My approach retains the metaphysical neutrality of the Finean formulation of grounding without resorting to his implausible appeal to the essences of the grounded. First, we are not committed to a grounding relation, as grounding claims need no grounding relations to be their truthmakers. Second, we are not committed to the existence of the grounded and the ground, but just to the concepts of them, which are already needed in our ontology if propositions are needed and they are compositional. Third, there is no commitment to reduction, but the grounded can be reduced to the ground as in the case of water is grounded in H 2 O. The relational concept of grounding holds between the concept of the ground and the concept of the grounded, which must be different concepts. But different concepts might have the same referent. In this way, the irreflexivity of grounding is secured without ruling out the possibility of reduction. In contrast, if we take grounding as a relation holding between metaphysically substantive facts, then we are practically ruling out the reduction of the grounded to the ground, given the irreflexivity of grounding. The implication of my approach is that the hierarchical structure of ontological dependence created by the grounding facts is at least con-

29 Jack Yip: Truthmaking and Grounding 27 ceptually substantive. Except for the concepts at the fundamental level, the set of non-fundamental concepts is not compulsory for a conceptual scheme. Instead of having the psychological, the biological, the chemical, and so on, a different set of such non-fundamental concepts can be made up and an alternative hierarchical structure can be built. So, it is just because our conceptual scheme is what it is that we have the hierarchical structure we have here lies the conceptual substantivity. But we can still conceive of the grounding facts as metaphysically substantive as concepts are part of the fundamental reality and the way they figure in the essence of one another is objective. Different sets of non-fundamental concepts can figure in the fundamental reality. So, even though there can be different hierarchical structures of ontological dependence generated by different sets of grounding facts, they can all be taken as metaphysically substantive. The hierarchical structure of ontological dependence exists because we do not usually conceive of the world as it is in itself. We have different ways to conceive of the same thing, and thereby come up with different concepts of the same thing. They can be related to one another by figuring in the essences of one another, and thus we can talk about the grounding relations between them. What is the concept of water but one of the results from our macroscopic way of conceiving of the world? This picture can justify metaphysicians endeavours to articulate the hierarchical structure of ontological dependence just by sitting on their armchairs and relying on the results of scientific discoveries. In their armchairs, metaphysicians have access to the concepts and they try to look into the essences of those concepts in order to see their dependence relations with one another (with the help of science). Assuming that physics is going to tell us what the fundamental ontological base (i.e. the non-representational part) is, metaphysicians aim at laying out the dependence relations between our non-fundamental concepts and the fundamental ones which match the structure at the base level. As in the whole time they are working with concepts, they do not have the need to get up from the armchair to fulfil their intellectual pursuits. 4 Conclusion If grounding is to be a useful notion to articulate ontological dependence, an account of how grounding facts can hold must be had. I have argued that no satisfactory account can be given under the framework of grounding. The notion of truthmaking can come in to help.

30 28 KRITERION Journal of Philosophy, 2015, 29(2): With a fundamental ontology of concepts as representational entities, we can have a satisfactory account of how grounding facts can hold by giving an answer to the question about what makes grounding claims true. As grounding claims are representational entities, grounding facts and subsequently the hierarchical structure of ontological dependence hold in virtue of something representational. The implication of this account is that what ontologically depends on what is a matter of our ways of conceptualising or representing the world, which means that it is a conceptually substantive matter. But given that concepts belong to the fundamental ontology, it is also a metaphysically substantive matter. Metaphysicians, sitting on their armchairs, can still play a role in articulating an area of reality for us. My account for how grounding facts hold amounts to having ungrounded grounding facts. The implausibility of this approach from the perspective of the grounding framework has been considered in the section 2. But this approach is still plausible, as we ve seen, if we can account for the holding of the ungrounded grounding facts within a framework of truthmaking. The dilemma between grounded and ungrounded grounding facts is a dilemma only if what we have in our hands is just grounding. But if we also have truthmaking, then one of the two options wins. The lesson is just that grounding is not going to accomplish everything, it might need help from other similar notions as well. But the help comes with important implications about what grounding is supposed to do. If some grounding theorist is not happy with them, he/she will have to find some other solution to the problem of how grounding facts hold. Notes 1 It s easy to see the inadequacy of modal notions: (1) Necessarily, the singleton of Socrates exists iff Socrates exists. (2) Assume that an entity A is ontologically dependent upon another entity B iff necessarily, if A exists, then B exists. Then, (3) Socrates is ontologically dependent upon the singleton of Socrates, and (4) the singleton of Socrates is ontologically dependent upon Socrates. But (3) is false. Perhaps we can adopt (2*) instead of (2): (2*) Assume that an entity A is ontologically dependent upon another entity B iff necessarily, if A exists, then B exists, but not vice versa. But then both (3) and (4) would be false. 2 Other formulations of grounding by a relational predicate can be derived from this canonical formulation. 3 Grounding facts is a suitable term for the second formulation, while grounding claims is more appropriate for the first formulation. But the problem of what grounds them should not hinge on the terminology.

31 Jack Yip: Truthmaking and Grounding 29 4 For the sake of exposition, I will take grounding as a relation and adopt a liberal view concerning the relata of a grounding relation. 5 To take grounding facts as ungrounded amounts to taking them as primitive. But this has to be distinguished from the primitiveness of the notion of grounding. 6 Kit Fine ([7]; [6]) uses the term reality instead, but I will stick to fundamentality. 7 Thanks to the suggestion of an anonymous referee. 8 Still, one might say that even if grounding facts are atomic, they might not be simple (as noted by Lewis in the case of structural universals [14]). As an anonymous referee points out, this case about fact composition is analogous to singletons of mereologically complex entities in Lewis s Megethology ([15]). But, even though such singletons are singletons of mereologically complex entities, the mereological complexity is disregarded in the composition of classes (which are fusions of singletons). By analogy, if grounding facts are atomic, even if they are complex in such a way to link up the grounded and the ground involved, this complexity should be disregarded in the composition of facts. Also, Lewis himself argues against a conception of simple which is distinguished from atomic ([14], p.41). 9 A possible example is an ontology of Aristotelian universals and thick particulars as fundamental entities. 10 As pointed out by an anonymous referee, this list is not logically exhaustive. But I believe that there are no other options that are not susceptible to arguments similar to the ones that follow. 11 Note that this regress is not vicious. 12 The involvement of other entities in the essence of a certain entity follows from Kit Fine s conception of essences ([5]). A real definition of an entity articulates the essence of an entity. It can be taken as a collection of propositions, which contains objects as constituents. There are two kinds of essences: constitutive and consequential. Here our focus is on constitutive essence which tells us what the entity is. A constitutive essence is a sub-group of propositions among the collection that are not the results of logical consequence from other propositions in the collection. Otherwise, it is a consequential essence. In our example, for P to ground G, G must belong either to the constitutive essence or to the consequential essence of P. Given the non-fundamental component L in G, not matter which kind G belongs to, the constitutive essence of P would involve some non-fundamental entity. If G belongs to the constitutive essence, then the constitutive essence of P would involve G, and thus L. If G belongs to the consequential essence, the constitutive essence of P would still involve something non-fundamental, because without any involvement of L or some city in the constitutive essence there is no way that a proposition involving G can be logically derived. 13 This regress might be vicious, as going through the chain, we have an infinity of the essences of different grounding facts. 14 Armstrong writes, To demand truthmakers for particular truths is to accept a realist theory for these truths. There is something that exists in reality, independent of the proposition in question, which makes the truth true. ([2], p.5) 15 I assume for the sake of exposition that propositions are truthbearers.

32 30 KRITERION Journal of Philosophy, 2015, 29(2): Many philosophers consider truthmaking and grounding as basically the same thing. They take truthmaking either as a restricted version of grounding ([7]) or as a species of grounding among others ([10]; [17]). But I have argued elsewhere that truthmaking and grounding should be two distinct notions ([22]). The principal reason is that the two notions have different theoretical aims: grounding articulates ontological dependences, while truthmaking articulates why the representational is true in virtue of the non-representational. 17 The fact here means just how things are. 18 For example, all crows are black and all non-black things are non-crows will be the same set of possible worlds. 19 For example, the truthmaker for there is a man cannot be picked out in this way as it is true in many different ways (i.e. being made true by different men). 20 In the case of the sentence water exists which contains the term water, we understand the same sentence with or without the knowledge that water is H 2 O. Similarly, in the case of the proposition water exists which contains the concept of water, we understand the same proposition with or without the knowledge that water is H 2 O. 21 These three terms are just terminological variants of the same thing ([13]; [6]; [19]; [21]). I will stick to the term fundamental. 22 For a concept A to figure in a concept B is for B s constitutive essence to contain the concept A (see also n.12). 23 We can account for the holding of the fact that the concept of H 2 O figures in the concept of water by considering again what makes the concept of H 2 O figures in the concept of water true. Jack Yip Department of Philosophy Room 10.13, 10/F Run Run Shaw Tower Centennial Campus The University of Hong Kong Pokfulam Road, Hong Kong <jackyip01@gmail.com> < References [1] D.M. Armstrong. A world of states of affairs. Cambridge University Press, Cambridge, 1997.

33 Jack Yip: Truthmaking and Grounding 31 [2] D.M. Armstrong. Truth and Truthmakers. Cambridge University Press, New York, [3] Louis derosset. Grounding Explanations. Philosophers Imprint 13(7):pp.1-26, [4] Kit Fine. Essence and Modality. Philosophical Perspectives 8:pp.1-16, [5] Kit Fine. Ontological Dependence, Proceedings of the Aristotelian Society 95:pp , [6] Kit Fine. The Question of Realism. Philosophers Imprint 1(1):pp.1-30, [7] Kit Fine. Guide to Ground. In Fabrice Correia and Benjamin Schnieder, editors, Metaphysical Grounding, pages pp Cambridge University Press, New York, [8] Gottlob Frege. Sense and Reference. Philosophical Review 57(3):pp , [9] Gottlob Frege. The thought: A logical inquiry. Mind 65(259):pp , [10] Aaron M. Griffith. Truthmaking and Grounding. Inquiry 57(2):pp , [11] C.S. Jenkins. Is Metaphysical Dependence Irreflexive?. The Monist 94(2):pp , [12] Saul Kripke. Naming and Necessity, Cambridge, Massachusetts, Harvard University Press, [13] David Lewis. New work for a theory of universals. Australasian Journal of Philosophy 61 (December):pp , [14] David Lewis. Against structural universals. Australasian Journal of Philosophy 64(1):pp , [15] David Lewis. Mathematics in Megethology. Philosophy Mathematica 1(1):pp. 3-23, [16] Hilary Putnam. The meaning of meaning. Minnesota Studies in the Philosophy of Science 7:pp , 1975.

34 32 KRITERION Journal of Philosophy, 2015, 29(2): [17] Gonzalo Rodriguez-Pereyra. Grounding is Not a Strict Order. Journal of the American Philosophical Association (forthcoming). [18] Gideon Rosen. Metaphysical Dependence: Grounding and Reduction. In Bob Hale and Aviv Hoffman, editors, Modality: Metaphysics, Logic and Epistemology, pages pp Oxford University Press, Oxford, [19] Jonathan Schaffer. On What Grounds What. In David Manley, David J. Chalmers & Ryan Wasserman, editors, Metametaphysics: New Essays on the Foundations of Ontology, pages pp Oxford University Press, New York, [20] Jonathan Schaffer. Grounding, Transitivity and Contrastivity. In Fabrice Correia and Benjamin Schnieder, editors, Metaphysical Grounding, pages pp Cambridge University Press, New York, [21] Theodore Sider. Writing the Book of the World. Oxford University Press, New York, [22] Jack Yip. Why Truthmaking Is Not Grounding. Presented in the 103rd Academic Year CTCT Postgraduate Philosophy Conference, Taipei, Taiwan, May [23] Stephen Yablo. Aboutness. Princeton University Press, New Jersey, 2014.

35 Metaontological Deflationism in the Aftermath of the Quine-Carnap Debate Jonathan Egeland Harouny Abstract With metaphysical philosophy gaining prominence in the aftermath of the Quine-Carnap debate, not only has it become assumed that the Quinean critique leaves ontological pluralism behind as an untenable approach, but also that the same is true of deflationism more generally. Building on Quine s criticisms against the analytic-synthetic distinction and the notion of quantifier variance, contemporary metaphysicians like van Inwagen and Sider continue to argue for the untenability of deflationary approaches to metaontology. In this paper I will argue that Quine s criticisms do not provide sufficient grounds for revitalizing metaphysics, as the aforementioned metaphysicians conceive them as doing, and that they also don t eliminate all hope for Carnapian pluralism. Furthermore, Carnap s initial position may even yield the most promising route for the pluralistically inclined. Moreover, pluralism is often conceived as being equivalent with the narrower notion of quantifier variance, often associated with Hirsch and Putnam. As this notion often is attributed not only to Carnap and other pluralists, but also is taken to be an essential feature of deflationism, explicating how their merits in fact don t necessarily coincide with those of quantifier variance will clarify matters. I will conclude by noting how neither pluralism nor deflationism is committed to quantifier variance, and thus how arguments against the latter don t entail a refutation of the former. Keywords: deflationism, Quine-Carnap Debate, Ontological Pluralism, Quantifier Variance 1 Introduction With metaphysical philosophy gaining prominence in the aftermath of the Quine-Carnap debate, not only has it become assumed that the Quinean critique leaves ontological pluralism behind as an untenable approach, but also that the same is true of deflationism more generally. Kriterion Journal of Philosophy, 2015, 29(2): c 2015 The author

36 34 KRITERION Journal of Philosophy, 2015, 29(2): Building on Quine s criticisms against the analytic-synthetic distinction and the notion of quantifier variance, contemporary metaphysicians like van Inwagen and Sider continue to argue for the untenability of deflationary approaches to metaontology. In this paper I will argue that Quine s criticisms do not provide sufficient grounds for revitalizing metaphysics, as the aforementioned metaphysicians conceive them as doing, and that they also don t eliminate all hope for Carnapian pluralism. Furthermore, Carnap s initial position may even yield the most promising route for the pluralistically inclined. As the most prominent deflationary approach, my arguments will employ ontological pluralism as the default deflationary position. But despite the close associations between pluralism and deflationism, where the latter often is conceived as being dependent on the former, deflationary approaches do not require pluralism as a necessary prerequisite, thus opening for the possibility that the positions might suffer different philosophical fates. Moreover, pluralism is often conceived as being equivalent with the narrower notion of quantifier variance, often associated with Hirsch and Putnam. As this notion often is attributed not only to Carnap and other pluralists, but also is taken to be an essential feature of deflationism, 1 explicating how their merits in fact don t necessarily coincide with those of quantifier variance will clarify matters. I will conclude by noting how neither pluralism nor deflationism is committed to quantifier variance, and thus how arguments against the latter don t entail a refutation of the former. 2 Carnap s Deflation of Metaphysics In order to see how Carnap deflates the semantics of metaphysical questions and discourse, we should start, as he himself does, by introducing the notion of linguistic frameworks. The introduction aims at squaring his empiricism with the acceptance of abstract objects without committing him to any metaphysical position, be it either platonism or nominalism. 2 The thesis here propounded states that the intelligibility of ontological questions is dependent on the questions being asked within a linguistic framework. Such a framework is a formal language, with semantic rules for forming and evaluating statements. The semantic rules operational within a given linguistic framework determine the range of meaningful sentence formation (involving a certain kind of entities) and assign truth-conditions to its sentences. Of the sentences and expressions involving terms and predicates that refer or apply to a certain kind of

37 Jonathan Egeland Harouny: Metaontological Deflationism 35 entities, the semantic rules further specify the observational conditions which count as confirming or disconfirming them. 3 By conceiving the semantics of existence-questions as determined by their overall relation to linguistic frameworks, Carnap posits a fundamental distinction between two kinds of questions concerning the existence or reality of entities ([3, p.17]). The distinction famously marks the difference between internal and external questions; 4 the former occurring within the framework, while the latter concern the existence or reality of the system of entities as a whole ([3, p.17]). Although formally presented as a dichotomy, I would argue that Carnap s conception of existence-questions can be analyzed in a more fruitful manner as being trichotomized. 5 Referring to a certain kind of entities, existence-questions can be understood in three ways. Firstly, they can be understood as internal questions, in which case their answers are provided by the semantic rules governing the use of the terms which refer to the postulated objects in question. If the postulated objects are concrete, their meaning, in accordance with the semantic rules governing the linguistic framework employed, specify the evidence that would empirically confirm or disconfirm statements attributing, e.g., the predicate apple to physical objects. If, on the other hand, the postulated objects are abstract, as in the case of numbers, then their meaning, which is found by logical analysis based on the rules for the new [numerical] expressions, alone determines the answer. Thus, answers to internal questions will be, depending on the nature of the objects in question, either determined by scientific-empirical investigations, or, as in the case of numbers, analytic. If the question rather is external, it might be so in two ways. It might be understood as factual-external, in which case its purport is to ask an ontological question about, e.g., numbers prior to the adoption of any framework. But, as Carnap made explicit with his notion of linguistic framework, speaking of a certain kind of entities implies a system of certain ways of speaking, subject to certain semantic rules of use. And in meaningfully asking if there really are numbers, one has already adopted the framework of numbers. Unless the philosopher purporting to ask a factual-external question is willing to recognize that his question really is of an internal kind with a trivial answer, he has not provided any cognitive content for it and so it must be regarded as a pseudo-question. As such, factual-external questions ought to be reinterpreted as pragmatic-external questions about which linguistic frameworks we ought to adopt. 6 As Carnap says in the case of such pseudo-questions:

38 36 KRITERION Journal of Philosophy, 2015, 29(2): [They are] disguised in the form of theoretical question[s] while in fact [they are] non-theoretical; in the present case it is the practical problem whether or not to incorporate into the language the new linguistic forms which constitute the framework of numbers. ([3, p.19]) As is stated above, the semantic, framework-internal rules governing the use of, e.g., number words, analytically entail trivial answers to arithmetical questions and to existence-questions concerning numbers. But how, one might ask, can the introduction of a linguistic framework yield definitive answers to questions of ontology? Carnap s answer is rather straightforward: the analyticity which makes internal questions and answers trivial places no constrains on reality and thus does not really make any (ontological) claims about it. The introduction of the new ways of speaking does not need any theoretical justification because it does not imply any assertion of reality. ([3, p.21]) But still commentators have found it plausible to regard the employment of different linguistic frameworks to entail commitment to different ontological entities. While some, e.g. Echehart Köhler ([11]), have taken Carnap s employment of the framework of numbers to commit him to platonism, others, e.g. Andre Gallois ([7]), consider the possibility of him really being a nominalist. But in placing the discussion between platonism and nominalism within the trichotomized version of the I/E distinction drawn above, we can easily see how the discussion really either is trivial, nonsensical, or pragmatic, depending on how it is framed. 7 To further explicate how the I/E distinction deflates metaphysics, we may describe it in terms of the use/mention distinction, 8 which clearly shows how existence-questions are linguistically framed. Using certain material terms, such as number and physical object, meaningfully presupposes conformity to the semantic framework-internal rules which both constitute legitimate use and deflate their alleged ontological status. On the other hand, as external questions don t conform to these rules they only succeed in mentioning such terms; which only suffices to serve a substantial function if they are posed as pragmatic-external questions. In light of Carnap s notion of linguistic frameworks, coupled with the I/E and the U/M distinctions, it becomes clear that he operates with a functional version of ontological pluralism. As different frameworks speak of different kinds of entity, they only do so in a functional and nonmetaphysical sense. Hence the reason why his pluralism construes ontological questions as really being of a linguistic kind. Carnap is neither saying that ontological objects of any particular kind do

39 Jonathan Egeland Harouny: Metaontological Deflationism 37 exist nor that they don t exist; his metaontology is deflationary through and through and he only concerns himself with ways in which we talk about ontological objects. In coupling his ontological pluralism and his deflationary attitude towards metaphysics, he moves away from the traditional realism/anti-realism dichotomy in the construal of a broader, non-metaphysical position. [I reject] both the thesis of the reality of the external world and the thesis of its irreality as pseudo-statements; the same [is] the case for both the thesis of the reality of universals and the nominalistic thesis that they are not real and that their alleged names are not names of anything but merely flatus vocis. (It is obvious that the apparent negation of a pseudostatement also must be a pseudo-statement). ([3, pp.21-22]) In the next section I will begin my discussion of Quine s critique of Carnap s metaontology and if it vindicates metaphysical philosophy, as we later will see some prominent metaphysicians believe it to do. The section s main focus will be on Quine s attack on the notion of analyticity. 3 Quine and the Problem of Analyticity Covering the first four parts of Quine s [18] article we find the circle argument, which purports to show how the notions of necessity and apriority, which Carnap s use of analyticity is supposed to explain, 9 really is presupposed by the analytic/synthetic distinction. 10 His claim is that attempts to define the notion analyticity by using terms thought to be synonymous or explicating, wind up employing terms without any greater intelligibility. Quine argues this to be the case when investigating formal definitions, cognitive synonymy, and semantical rules, none of which are able to provide any further understanding of analyticity without circularly appealing to the notion of analyticity in the explanation; in some cases the terms thought to be synonymous are intelligible only if the notion of analyticity is already clearly understood in advance ([18, p.30]). In attempting to save analyticity and synonymy from the Quinean critique, Carnap, by arguing that intension is scientifically verifiable and legitimate, attempts to show how the meanings of coextensive predicates in principle are empirically distinguishable. Thus, by appeal to a scientifically respectable notion of meaning, Carnap argues that we can define the concepts of synonymy and analyticity; and so he does:

40 38 KRITERION Journal of Philosophy, 2015, 29(2): Two expressions are synonymous in the language L for X at a time t if they have the same intension in L for X at t. A sentence is analytic in L for X at t if its intension [... ] in L for X at t comprehends all possible cases. ([2, p.42]) But intuitively, Quine s initial objection still seems strong, even against this formal definition of analyticity. Here I agree with the status of the situation as Scott Soames ([25, p.437]) puts it: namely, that it is far from clear that Carnap escapes the circle argument as his defense of intension rests heavily on modal claims about what a predicate would apply to [... ] were certain possible circumstances to obtain. 11 Now, if Carnap was to sever the I/E distinction from the A/S distinction by throwing this broad conception of analyticity over board, the problem remains that even if trivial answers to existence-questions are somehow ( analytically ) entailed by their meanings, their status as metaphysically innocent is not. The reason for this is that if (the modal notion of) necessity is abandoned when leaving the broad conception of analyticity presupposed by the I/E distinction, one can never be sure that ontological questions/answers fall neatly into either side of the distinction. And this is the exact point conveyed by Quine s gradualism, as he envisages there being a continuum of existence questions, where the differences between them only are in degree and not in kind ([17, p.72]). But where does Quine s merging of existence-questions leave metaphysics? 12 As Quine later in his Word and Object makes explicit, he merges the opposites of Carnap s (I/E) model by maintaining that all theoretical questions in general are of a pragmatic nature, hence rendering the category of purely internal questions empty ([22, p.271]). In drawing the conclusions of pragmatism further than Carnap did, the main effect of Quine s argument is that questions of scientific and mathematical matters lose some of their logical purity, and not that metaphysical questions experience any such gain. The point missed by metaphysicians gladly quoting Quine when he states that the conflation of the A/S distinction implies that ontological questions then end up on a par with questions of natural science ([17, p.71]), is that he does not conceive questions of metaphysics to regain the status of empirical investigations as traditionally conceived (i.e., as theoretical), but that empirical, scientific and mathematical questions are like Carnap s conception of metaphysical questions (i.e., pragmatic). Thus, in the spirit of what Huw Price ([12, p.327]) has called a more thoroughgoing post-positivist pragmatism, Quine, rather than vindicating traditional metaphysics, rather expands the boundaries of what Carnap s deflationism initially allowed: Carnap

41 Jonathan Egeland Harouny: Metaontological Deflationism 39 maintains that ontological questions [... ] are questions not of fact but of choosing a convenient conceptual scheme or framework for science; and with this I agree only if the same be conceded for every scientific hypothesis. ([17, p.72]) 4 Is Carnap s Ontological Pluralism Tenable? Quine s Charge and Ryle s Defense If Quine s critique of the A/S distinction doesn t only blur the lines that separate theoretical from pragmatic questions, but also implies that framework-internal questions are indistinguishable from pragmaticexternal questions, what is to stop us from treating all questions as internal to a single grand framework? And this is precisely Quine s second charge against Carnap s (I/E) model. Whereas one of the purposes of Carnap s model is allowing theories committed to abstract entities to feature as metaphysically innocent ([3, pp.16-17]), the point made by Quine s famous slogan to be assumed as an entity is [... ] to be reckoned as the value of a variable ([16, p.12]), is not that the reality of what exists is reduced to being the value of a bound variable, but that to be ontologically committed to the existence of something is to assert the existentially quantified claim that there are entities of the asserted kind. Hence, while Carnap believes his commitment to numbers is unworthy of ontological worry, Quine makes explicit how such a commitment entails platonism ([16, p.13]). By conceiving Carnap s linguistic frameworks as continuous language fragments within a single grand framework in which a single univocal quantifier ranges, Quine claims the fact that quantification over mathematical entities is indispensable to our best scientific theories ought to commit us to the ontological existence of such entities. 13 Continuing, let s turn to how exactly Quine purports to show that we only operate with a single existential quantifier that, rather than only being operational within a distinct domain, is allowed to range over anything. In recasting Carnap s I/E distinction as derivative from the more fundamental distinction between category- and subclass questions 14 ([17, pp.68-69]), Quine argues that the introduction of distinct frameworks into language is dependent on the rather trivial consideration of whether we use one style of variables or [several] ([17, p.69]). To use his own example: Whether the statement that there are physical objects and the statement that there are black swans should be put on the same side of the dichotomy, or on opposite sides, ([17, p.69]) is completely

42 40 KRITERION Journal of Philosophy, 2015, 29(2): determined by arbitrary choices of quantificational apparatus. By homogenizing the existential quantifier, so that various kinds of object are expressed by the same kind of variables, even general questions of existence and metaphysics may be allowed the status as subclass. Thus, by homogenizing the existential quantifier, Quine again threatens the deflationism entailed by Carnap s model. Whereas we saw his rejection of the A/S distinction far from worsening the deflationist s claim, rather advancing its initial pragmatic concerns to cover new ground, Quine s attack on Carnap s pluralism rejects the idea that different language fragments are semantically isolable in virtue of the quantifiers they employ. As Carnap doesn t provide a principled method for distinguishing subclass questions from category questions, if he is to recover his position and fend of Quine s charge against his pluralism, he somehow has to be able to turn the weakness of his model into its strength. In other words, he has to maintain his initial position meanwhile denying that there aren t any boundaries to be marked regarding existence-questions prior to our quantificational choices. I believe the most promising option for recovering Carnapian deflationism and removing the sting from Quine s criticism, is to argue that the pluralism in question is determined by other aspects of linguistic functionality rather than being a matter of arbitrary quantificational choice: namely, that his pluralism is of a functional nature in that the differences (of kind) between existencequestions (e.g., about whether there are any apples in my kitchen, or whether natural numbers exist,) makes the quantificational pluralism expressed in his model determined by this more basic linguistic fact. 15 As recently suggested by Price, 16 this sort of defense has been advanced by Gilbert Ryle in The Concept of Mind. The pluralism here endorsed makes explicit that we may allow our existential quantifier to be homogenized, as Quine suggests, while maintaining that the same concept of existence plays different functional roles in our language. It is perfectly proper to say, in one logical tone of voice, that there exist minds and to say, in another logical tone of voice, that there exist bodies. But these expressions do not indicate two different species of existence, for existence is not a generic word like coloured or sexed. They indicate two different senses of exist, somewhat as rising has different senses in the tide is rising, hopes are rising, and the average age of death is rising. ([23, p.12]) The insight here provided by Ryle may allow our (singular) concept of existence to feature both in talk of concrete and abstract objects,

43 Jonathan Egeland Harouny: Metaontological Deflationism 41 while recognizing that the difference in admitting the existence of both kinds of entity is due to the difference in talk about, e.g., apples and numbers, rather than using different concepts of existence. Ryle s functional version of ontological pluralism, rather than focusing on what is talked about in differently framed discourses, is more sensitive to the function a given framework serves and what its talk is for. 17 With this in mind, even by discarding the I/E distinction à la Quine, the Carnapian who accepts Ryle s argument might be able to account for different quantifiers that feature in different parts of language, if the different quantifiers can be understood as being determined by the functional character of the language fragments in which they occur. And in this way the Carnapian position may be tenable and a more modest notion of linguistic framework may be reintroduced, even if we take Quine s criticisms into account; for, as we have seen, it is possible to differentiate between existence-questions prior to quantificational choices. To summarize, Ryle s argument may be taken to support an account of existential quantification where quantificational semantics is functionally determined by other parts of language. If we understand Carnap as entertaining the idea of different existence-concepts capable of various senses, better known as quantifier variance, we may say that his frameworks introduce different senses of exists, such as exists q, exists r, exists s, and so on, where q xf x will be true in a linguistic framework I if xf x is true at the furnished framework I, f(i), where f is a certain furnishing function. (The Carnapian should, of course, provide a plausible account of the furnishing function without appeal to analyticity.) But it should be noted that his pluralism doesn t seem to commit him to quantifier variance. The introduction of a new material concept into language does not entail any significant change in the rules of use governing the formal existence concept. Now, of course, the introduction of a new material concept may increase the domain in which the existential quantifier is allowed to range and thus effect a rather minimal change in its meaning, but it does not follow that the general rules of use for the quantifier radically differ from one linguistic framework to the other, as is supposed by quantifier variance. Furthermore, as Amie Thomasson has argued rather plausibly ([27]; Cf. [26, pp.69-79,159]), Carnap does not conceive the nature of ontological debates as verbal because quantifiers vary between different frameworks, but because the different parties of the debate fall on different sides of the U/M distinction with regard to the relevant material concepts being debated. The major concern of Carnap s pluralism

44 42 KRITERION Journal of Philosophy, 2015, 29(2): does not pertain to the minor changes in how our use of the existence concept varies between different frameworks say the nominalist and the platonist framework but rather how the latter accepts and makes use of material concepts with reference to abstract objects, while the former, insofar as he wants to say something substantial about the material concepts in question, only may mention them. On this reading the nominalist cannot consistently make use of the concept of number meanwhile denying the existence of numbers; for in using the concept he also has to assent to the existence of numbers, as the rules of use governing the concept entail. 18 On the other hand, I take Ryle s functional version of ontological pluralism to be committed to quantifier variance; the view that the sense of existence may vary. Both accounts of existential quantification are of course limited to lightweight quantification, meaning that they regard answers to questions of ontology as trivial, while discarding the absolute quantifier as meaningless. In this section I have shown that Ryle s argument, in making the pluralistic semantics of exists functionally determined prior to quantificational choices, may be appropriated by Carnapianism in response to Quine s criticism. Whichever route is preferable, also of how to understand Carnap s pluralism, as for now, both remain available for deflationism. In the next section I will consider some common objections to ontological pluralism and possible answers. As will become clear, the importance of pluralism s role in deflationism should be of great concern, and providing certain remarks on the issue will become necessary. 5 The Criticisms of Ontological Pluralism: Still a Viable Candidate for Deflationism? The first problems to be noted should be the ones that loom large over my representation of Ryle s pluralistic conception of language-use: namely, that we ve neither provided any criteria for the presence of functional differences in our language, nor for what constitutes these differences. These problems do not only require further philosophical reflection, but, more importantly, also rigorous empirical investigations of how we use language. I will therefore put this problem aside as one worthy of more detailed discussions and further empirical studies, while moving on to other difficulties the pluralist faces. As ontological pluralism often goes under the guise of (one of its members,) quantifier variance, 19 that is where we ll begin. In recent discussions the most prominent names associated with quantifier vari-

45 Jonathan Egeland Harouny: Metaontological Deflationism 43 ance have come to be Hilary Putnam and Eli Hirsch. As proponents of the doctrine, they both agree that existence may be expressed by different, but logico-syntactically similar, 20 concepts: [The] phenomenon [of conceptual relativity/quantifier variance] turns on the fact that the logical primitives themselves, and in particular the notions of object and existence, have a multitude of different uses rather than one absolute meaning. ([13, p.19]; [15, p.37,43]; Cf. [8]; [10].) For present purposes we ll sweep their differing conceptions of quantifier variance under the rug 21 and focus on some of the criticisms their position(s) has generated. By using Peter van Inwagen s criticisms as an example, it will become clear how the point of contention of metaphysicians defending heavyweight quantification 22 is that there is only one univocal, existential quantifier and that its meaning is invariant. In drawing analogies between our use of number and our use of existence, van Inwagen suggests that the univocacy of number and the intimate connection between number and existence should convince us that there is at least very good reason to think that existence is univocal ([30, p.482]). Continuing, he further claims that the operator there exists is intimately related to disjunction [and all ] ([30, p.484]). He gives us the following example: There exists a prime number between 16 and 20 is equivalent to 17 is a prime or 18 is prime or 19 is prime only given that 17, 18, and 19 are all the numbers between 16 and 20 ([30, p.484]). Due to the univocal all and its interdefinability with exists, the argument purports to show that exists also is univocal. As mentioned, the point of the arguments of van Inwagen and other metaphysicians seems to be that we only operate with a single concept of existence and that its meaning is invariant. Intuitively, this and other arguments against quantifier variance seem to plausibly account for our logical vocabulary as not being vague and varying in meaning, but in that Carnap needn t be interpreted as a quantifier variantist, it is not clear that his deflationism loses its legitimacy. Before I consider another objection to ontological pluralism, I should note that I believe one of the reasons why the metaphysical branch of philosophy emerged victorious from the Quine-Carnap debate is that criticisms of the kind above are falsely taken not only to count against Carnap, but also more generally to refute ontological pluralism and deflationism. The reason for all this seems to be that Carnapianism, pluralism and deflationism often are conceived as entailing and relying on quantifier variance. As Ted Sider remarks: The deflationist must claim that the participants in ontolog-

46 44 KRITERION Journal of Philosophy, 2015, 29(2): ical debates mean different things by the quantifiers. And so, the deflationist must accept that quantifiers can mean different things, that there are multiple candidate meanings for quantifiers. In Hirsch s phrase, deflationists must accept quantifier variance. ([24, p.391]; Cf. [10]) As we have seen above, Carnap most likely does not need to commit himself to the doctrine, despite philosophers like Matti Eklund ([4, p.137]) and Kit Fine ([6, p.164, fn2]) explicitly attributing it to him. Such considerations should not only make it obvious that ontological pluralism doesn t depend on quantifier variance, but also, and even more importantly, that deflationism doesn t. So, when prominent contemporary metaphysicians like Sider argue against deflationism and ontological pluralism by appealing to the weaknesses of quantifier variance, we should not take this to provide sufficient reasons for refuting such positions, but rather as halting one of their most recent forms in its track. Before I conclude with a short note on the role of pluralism in deflationism, I will consider one last criticism of pluralism forwarded by Eklund in a recent paper ([4]). Here the problematic nature of ontological pluralism is made explicit by how he defines the position; which really is Hirsch s version of ontological pluralism: 23 What any ontological pluralist view involves is roughly, see immediately below the following: There are a number of different languages we could speak, such that (a) different existence sentences come out true in these languages, due to the fact that the ontological expressions (counterparts of there is, exists, etc.) in these languages express different concepts of existence, and (b) these languages can somehow describe the world s facts equally well and fully (maybe some of these languages are more convenient to use than others but that is a different matter). ([4, p.137]) As I have already dealt with the first part of the definition, (a), under the name of quantifier variance, I ll be focusing on the second part, (b). Eklund s reason for including it, even though he admits that ontological pluralists tend not to explicitly introduce a condition like this when they describe their doctrine ([4, p.139]), is that he conceives it as needed. 24 If I understand his reasoning correct, Eklund believes this to be the case because we cannot make adequate sense of a language with more/less expressive powers than our own. This he further takes to commit the pluralist to the counterintuitive view that with respect to

47 Jonathan Egeland Harouny: Metaontological Deflationism 45 all subject matters [... ] there will be equally good theories formulated in these possible equally good languages ([4, p.140]). 25 As will be obvious to any reader, the reason why pluralists often don t explicate any such condition is that its anything goes -attitude seems, at worst, untenable, and, at best, quite unsettling. But I do not think the pluralist needs to be committed to any such condition. If the condition (b) under discussion means to say that these languages can somehow describe the world s [metaphysical] facts equally well and fully, then it is clear that the pluralist doesn t require appealing to it. For Carnap, being such a pluralist, may say that different linguistic frameworks do not describe the world s [metaphysical] facts at all, but rather concern their own furnished contents. Whether some frameworks come closer to describing these facts, or even if there aren t any such facts to somehow describe (idealism), it is evident from Carnap s position that such speculation, including the (b) condition, is senseless. This is evidently the case because it requires the adoption of a neutral position outside all frameworks, which would not conform to any semantic rules constitutive of meaningful sentence formation. If the (b) condition on the other hand not is taken as appealing to metaphysical correspondence 26 for providing a standard of judging different languages, it seems that the convenience of different languages, which Eklund explicitly takes to be another matter, provides us with the only possible standard for language choice. By excluding metaphysics from meaningful discourse the ontological pluralist may still insist on the utility of certain languages despite not operating with the notion of metaphysical correspondence, as for example Carnap, earlier noted, seems to do. 27 Thus, it will be clear how (less pragmatic) talk of water deprived of chemical sophistication at some point was traded for the (the more pragmatic) talk of water that furnished our concept with the property H 2 O, without having to consult the metaphysician in order to consider the latter of greater expressive power. Although ontological pluralism and even deflationism more generally often are conceived as necessarily sharing the fate of quantifier variance, they certainly do not. On the contrary, it does not only seem like both positions still are viable, but also that the initial position forwarded by Carnap ([3]) 28 still may yield one of the most prominent pluralist approaches to metaontology. Moreover, not only is pluralism still a viable candidate for deflationism, it is just one among several. Although I won t provide any detailed account of the positions here, other viable candidates may, e.g., be ontological maximalism which is the view

48 46 KRITERION Journal of Philosophy, 2015, 29(2): that anything that could be said to exist in any given language in fact does exist and radical fictionalism 29 according to which ontological statements always lack literal content. This is not to say the deflationist should accept Carnapian pluralism, or maximalism or fictionalism for that matter, but rather that he could do so. The tenability of deflationism in general does not depend exclusively on the merits of pluralism, and especially not those pertaining to quantifier variance. 6 Conclusion I have argued that Quine s critique of Carnap s deflationary approach to ontology, while weakening the initial position, does not leave it without possibility for recovery. The pluralistic approach so often criticized, despite the objections of metaphysicians like van Inwagen and Sider, may still find a tenable approach in Carnap s initial position forwarded in Empiricism, Semantics, and Ontology. But even if this position is found wanting, there are still other ways for the deflationist to go. We can learn from the Quine-Carnap debate, and the metaontological discussions in its aftermath, that despite metaphysics (unofficially) being crowned the victor, deflationism still provides viable approaches against the contemporary philosophical establishment. Notes 1 [24, p.391]. 2 [3, pp.16-17]. 3 [3, p.17]. 4 Continuing, the distinction will be abbreviated as the I/E distinction. 5 For a similar approach, see [4, pp ]. 6 See [3, pp.18-19] for Carnap s discussion on numbers. 7 The main point of the trichotomized version of the I/E distinction, as I have presented it here, is to show how metaphysical (i.e. factual-external) questions differ from pragmatic (i.e. pragmatic-external) questions. Although Carnap did believe that both kinds of question were devoid of cognitive content, only the latter kind is able to serve any substantial function at all: namely, the pragmatic function of evaluating the merits of linguistic frameworks not yet adopted. 8 Continuing, the distinction will be abbreviated as the U/M distinction. 9 Cf. [1, chapter 4]. 10 Continuing, the distinction will be abbreviated as the A/S distinction. 11 All logically possible cases come into consideration for the determination of intensions. ([2, p.38]).

49 Jonathan Egeland Harouny: Metaontological Deflationism After all he does suspect that the notion of such a [analytic/synthetic] dichotomy only encourages confused impressions of how language relates to the world. ([22, p.67]). 13 For more on how we ought to be ontologically committed to the entities which our best scientific theories quantify over, and especially entities of mathematics, see [16]; [19]; [20]; [21]; Cf. [14]. 14 The main difference to be noticed here is that internal questions don t only comprise subclass questions, but also category questions when these are asked in conformity to the semantic rules within the adopted framework, thus entailing either trivially analytic or contradictory answers ([17, p.69]). 15 Later in this section I will remark that Carnap s view, contrary to what I just stated, isn t committed to what I called quantificational pluralism, or what is better known as quantifier variance. The reason why I treat him as having such a commitment is that Carnap, ontological pluralism, and deflationism more generally often are seen as necessarily presupposing such a commitment. 16 [12, pp ]. 17 It is lack of such sensitivity to linguistic functionality, as is manifest in metaphysical philosophy, he believes to cause category mistakes i.e., mistakes in which something belonging to a particular category is described as if it belonged to a different category. 18 Cf. [27]. 19 As an example, Jason Turner ([28, pp.9-10]; [29, pp ]) even seems to assimilate the notions. 20 E.g. [8, p.53]. 21 The main difference, as I see it, is that whereas Putnam argued in favor of actual quantifier variance ([13, p.19]), Hirsch argues for the more modest position of possible quantifier variance ([8, p.60]). It should also be noted that whereas Putnam wrongly thought of conceptual relativity as incompatible with realism ([13, pp.35-36]), Hirsch in fact exposed how this assumption rested on a U/M mistake. (See [8, pp.52-53]). 22 The claim that there are answers to ontological questions and that they are nontrivial. 23 See [8, p.57]; Cf. [9, p.231], for Hirsch s agreement with (b). 24 In a more recent article [5] makes explicit that what is needed is an equivalence relation intermediate in strength between different theories formulated in different languages. 25 Since Eklund doesn t see Carnap as fulfilling this requirement, he sees him as an erring pluralist ([4, pp ]). 26 Many of the problems Eklund identifies with ontological pluralism and deflationism comes from the fact that he conceives them as metaphysical theses that picture reality as an amorphous lump ([5]). But, as I have attempted to show, the theses could be restricted to our use of ontological concepts and needn t have any metaphysical entailments. In other words, the pluralist might be said to endorse semantic vagueness, without having to endorse ontic vagueness. This view would also refute the claim of [5] that the pluralist should adopt maximalism; the thesis that anything which possibly could exist in fact does exist. A natural problem for the picture of reality as an amorphous lump, which counts against

50 48 KRITERION Journal of Philosophy, 2015, 29(2): deflationism as conceived by Eklund, is that incompatible or contradictory descriptions of the world in different languages might not leave any room left for the lump metaphor. For if, as I argue, the pluralist only needs to be able to differentiate between different and incompatible languages on pragmatic grounds, without having to make assumptions about any underlying reality, it seems that considerations of parsimony should leave out any metaphysical commitments. 27 Such a position could possibly also be compatible with a view on semantics that involves a rethinking of both the role and concept of reference in such a way that would eliminate the threat from what [5] calls the Tarskian argument and the sameness argument. 28 In modified form if interpreted as being committed to quantifier variance, as noted above. 29 I forward this position as a radicalization of the fictionalism endorsed by Stephen Yablo ([31]). According to Yablo s more modest position, instead of being ontologically committing, ontological statements might rather be of a make-believe or fictional nature. Jonathan Egeland Harouny University of Bergen <jonathaneh92@hotmail.com> References [1] Ayer, A. J. (1971) The A Priori. In: Language, Truth and Logic. Penguin. pp [2] Carnap, R. (1955) Meaning and Synonymy in Natural Languages. Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, 6, 3, pp [3] Carnap, R. (1956) Empiricism, Semantics, and Ontology. Reprinted in: Kim, J., Korman, D. Z. & Sosa, E. Ed. (2012) Metaphysics: An Anthology. Wiley-Blackwell. pp [4] Eklund, M. (2009) Carnap and Ontological Pluralism. In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp

51 Jonathan Egeland Harouny: Metaontological Deflationism 49 [5] Eklund, M. (2013) The Picture of Reality as an Amorphous Lump. In: Sider, T., Hawthorne, J. & Zimmerman, D. W. Ed. Contemporary Debates in Metaphysics. Wiley-Balckwell. pp [6] Fine, K. (2009) The Question of Ontology. In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp [7] Gallois, A. (2003) Does Ontology Rest on a Mistake? Aristotelian Society Supplementary Volume, 72, 1, June 1998, pp [8] Hirsch, E. (2002) Quantifier Variance and Realism. Philosophical Issues, 12, 1, pp [9] Hirsch, E. (2004) Sosa s Existential Relativism. In: Greco, J. Ed. Ernest Sosa: And His Critics. Blackwell Publishing. pp [10] Hirsch, E. (2013) Ontological Arguments: Interpretive Charity and Quantifier Variance. In: Sider, T., Hawthorne, J. & Zimmerman, D. W. Ed. Contemporary Debates in Metaphysics. Wiley- Balckwell. pp [11] Köhler, E. (2006) Ramey and the Vienna Circle on Logicism. In: Galavotti, M. Ed. Cambridge and Vienna: Frank P. Ramsey and the Vienna Circle. Springer. pp [12] Price, H. (2009) Metaphysics after Carnap: The Ghost Who Walks? In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp [13] Putnam, H. (1987) The Many Faces of Realism. Open Court Publishing Company. [14] Putnam, H. (1975) What is mathematical truth? In: Mathematics, Matter and Method: Philosophical Papers, Volume 1. London. Cambridge University Press. pp [15] Putnam, H. (2004) Lecture 2: A Defense of Conceptual Relativity. In: Ethics Without Ontology. Harvard University Press. pp [16] Quine, W. V. (1948) On What There Is. Reprinted in: Kim, J., Korman, D. Z. & Sosa, E. Ed. (2012) Metaphysics: An Anthology. Wiley- Blackwell. pp

52 50 KRITERION Journal of Philosophy, 2015, 29(2): [17] Quine, W. V. (1951a) On Carnap s View s on Ontology. Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, 2, 5, pp [18] Quine, W. V. (1951b) Two Dogmas of Empiricism. The Philosophical Review, 60, 1, p [19] Quine, W. V. (1960) Carnap and Logical Truth. Synthese, 12, 4, pp [20] Quine, W. V. (1981a) Things and Their Place in Theories. In: Theories and Things. Cambridge. Harvard University Press. pp [21] Quine, W. V. (1981b) Success and Limits of Mathematization. In: Theories and Things. Cambridge. Harvard University Press. pp [22] Quine, W. V. (2013) Ontic Decision. In: Word and Object. Martino Fine Books. [23] Ryle, G. (2009) Descartes Myth. In: The Concept of Mind. Routledge. pp [24] Sider, T. (2009) Ontological Realism. In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp [25] Soames, S. (2009) Ontology, Analyticity, and Meaning: The Quine- Carnap Dispute. In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp [26] Thomasson, A. L. (2014) Ontology Made Easy. Oxford University Press. [27] Thomasson, A. L. (forthcoming) Carnap and the Prospects for Easy Ontology. Forthcoming in: Blatti, S. & LaPointe, S. Ed. Ontology after Carnap. Oxford University Press. [28] Turner, J. (2010) Ontological Pluralism. Journal of Philosophy, 107, 1, pp [29] Turner, J. (2012) Logic and Ontological Pluralism. Journal of Philosophical Logic, 41, 2, pp

53 Jonathan Egeland Harouny: Metaontological Deflationism 51 [30] Van Inwagen, P. (2009) Being, Existence, and Ontological Commitment. In: Chalmers, D. J., Manley, D. & Wasserman, R. Ed. Metametaphysics: New Essays on the Foundations of Ontology. Oxford University Press. pp [31] Yablo, S. (1998) Does Ontology Rest on a Mistake? Reprinted in: Kim, J., Korman, D. Z. & Sosa, E. Ed. (2012) Metaphysics: An Anthology. Wiley- Blackwell. pp

54

55 Wittgensteinian Epistemology and Cartesian Skepticism Nicola Claudio Salvatore Abstract In this paper, I present and criticize a number of influential anti-skeptical strategies inspired by Wittgenstein s remarks on hinges. Furthermore, I argue that, following Wittgenstein s analogy between hinges and rules of grammar, we should be able to get rid of Cartesian skeptical scenarios as nonsensical, even if apparently intelligible, combinations of signs. Keywords: Wittgenstein, rules of grammar, skepticism 1 The Cartesian Skeptical Paradox The feature of Cartesian style arguments is that we cannot know certain empirical propositions (such as Human beings have bodies, or There are external objects ) as we may be dreaming, hallucinating, deceived by a demon or be brains in the vat (BIV), that is, disembodied brains floating in a vat, connected to supercomputers that stimulate us in just the same way that normal brains are stimulated when they perceive things in a normal way. 1 Therefore, as we are unable to refute these skeptical hypotheses, we are also unable to know propositions that we would otherwise accept as being true if we could rule out these scenarios. Cartesian arguments are extremely powerful as they rest on the Closure principle for knowledge. According to this principle, knowledge is closed under known entailment. Roughly speaking, this principle states that if an agent knows a proposition (e.g., that she has two hands), and competently deduces from this proposition a second proposition (e.g., that having hands entails that she is not a BIV), then she also knows the second proposition (that she is not a BIV). More formally: The Closure Principle If S knows that p, and S competently deduces from p that q, Kriterion Journal of Philosophy, 2015, 29(2): c 2015 The author

56 54 KRITERION Journal of Philosophy, 2015, 29(2): thereby coming to believe that q on this basis, while retaining her knowledge that p, then S knows that q 2. Let s take a skeptical hypothesis, SH, such as the BIV hypothesis mentioned above, and M, an empirical proposition such as Human beings have bodies that would entail the falsity of a skeptical hypothesis. We can then state the structure of Cartesian skeptical arguments as follows: (S1) I do not know not-sh (S2) If I do not know not-sh, then I do not know M (SC) I do not know M Considering that we can repeat this argument for each and every one of our empirical knowledge claims, the radical skeptical consequence we can draw from this and similar arguments is that our knowledge is impossible. 2 Wittgenstein on Skepticism; A Minimal Reading A way of dealing with Cartesian style skepticism is to deny the premise S1) of the skeptical argument, thus affirming contra the skeptic that we can know the falsity of the relevant skeptical hypothesis. For instance, in his A defence of commonsense ([16], henceforth DCS) and Proof of the external world ([17], henceforth PEW), G. E. Moore famously argued that we can have knowledge of the commonsense view of the world, that is, of propositions such as, Human beings have bodies, There are external objects or The earth existed long before my birth and that this knowledge would offer a direct response against skeptical worries. Wittgenstein wrote the 676 remarks published posthumously as On Certainty ([48], henceforth OC) under the influence of DCS and PEW, and in particular in the context of conversations he had about these papers with his friend and pupil Norman Malcolm 3. As I have briefly mentioned supra, according to Moore, it is possible to provide a direct refutation of Cartesian-style skepticism, thus claiming contra the skeptic that we can know the denials of skeptical hypotheses. But, Wittgenstein argues, to say that we simply know Moore s obvious truisms is somewhat misleading, for a number of reasons. Firstly (OC 349, 483), because in order to say I know one should be able, at least in principle, to produce evidence or to offer compelling grounds for his beliefs; but Moore cannot ground his knowledge-claims

57 Nicola Claudio Salvatore: Wittgensteinian Epistemology 55 with evidence or reasons because (OC 245) his grounds aren t stronger than what they are supposed to justify. As Wittgenstein points out, if a piece of evidence has to count as compelling grounds for our belief in a certain proposition then that evidence must be at least as certain the belief itself. This cannot happen in the case of a Moorean commonsense certainty such as I have two hands because, at least in normal circumstances, nothing is more certain than the fact that we have two hands ([34]). As Wittgenstein writes in OC: If a blind man were to ask me Have you got two hands? I should not make sure by looking. If I were to have any doubt of it, then I don t know why I should trust my eyes. For why shouldn t I test my eyes by looking to find out whether I see my two hands? What should be tested by what? (OC 125). Imagine, for instance, that one attempted to legitimate one s claim to know that p by using the evidence that one has for p (for example, what one sees, what one has been told about p and so on). Now, if the evidence we adduced to support p was less secure than p itself, then this same evidence would be unable to support p: My having two hands is, in normal circumstances, as certain as anything that I could produce in evidence for it. That is why I am not in a position to take the sight of my hand as evidence for it (OC 250). Moreover, Wittgenstein argues, a knowledge-claim can be challenged by, for instance, the appeal to evidence and reasons; more generally, when we challenge a knowledge claim we can recognize what and if something has gone wrong in the agent s process of knowledge-acquisition. Things are somewhat different in the case of the denials of Moore s obvious truisms of the commonsense ; if, for instance, I believe that I am sitting in my room whilst I am not, there are no grounds on which this belief could be explained as a mistake, as an error based on negligence, fatigue or ignorance. On the contrary, a similar false belief would more likely be the result of a sensorial or mental disturbance (OC 526). As Moyal- Sharrock points out ([18], 74), in fact, for Wittgenstein if someone was holding seriously a denial of Moore s truisms (i.e., she believed she had no body or that both her parents were men) we would not investigate the truth-value of her affirmations, but instead her ability to understand the language she is using or her sanity (OC 155). If Moore s commonsense certainties are still not knowable, argues Wittgenstein, they are immune from rational doubt. This is so (OC

58 56 KRITERION Journal of Philosophy, 2015, 29(2): ) because doubts must be based on grounds; that is, they have to be internal to a particular practice and must be in some way or another justified. If they aren t, they are constitutively empty. To illustrate this point, Wittgenstein gives the example (OC 310) of a pupil who constantly interrupts a lesson, questioning the existence of material objects or the meaning of words; far from being a legitimate intellectual task, the pupil s doubt will lack any sense and will at most lead to a sort of epistemic paralysis, for she will just be unable to learn the skill or the subject we are trying to teach her (OC 315). Accordingly, as per Wittgenstein, all reasonable doubts presuppose certainty (OC ); that is, the very fact that we usually raise doubts of every sort at the same time shows and implies that we take something for granted. For example, a doubt about the real existence of an historical figure presupposes that we consider certain an obvious truism of the commonsense such as, The world existed a long time before my birth ; a doubt about the existence of a planet presupposes the absence of any doubt about the existence of the external world and so on (OC , ). But if the statements listed by Moore in DCS are not knowable or doubtable, what is their status? With regard to Moore s truisms, Wittgenstein introduces a concept that is pivotal to understand his antiskeptical strategy and at the same time extremely elusive: Moore s commonsense certainties are, in his words, hinges. Wittgenstein uses this term on different occasions, as in OC 341-3, where he writes: The question that we raise and our doubts depend on the fact that some propositions are exempt from doubt, are as it were the hinges on which those turn [....] that is to say, it belongs to the logic of our scientific investigations that certain things are in deed not doubted [... ] If I want the door to turn, the hinges must stay put. That is to say, hinges are just apparently empirical contingent claims; on closer inspection, they perform a different, more basic role in our epistemic practices. 3 Wright s Unearned Warrant So far, I have just sketched Wittgenstein s anti-skeptical reflections. Given the elusiveness and obscurity of his work, there is no consensus

59 Nicola Claudio Salvatore: Wittgensteinian Epistemology 57 on how we should interpret Wittgenstein s anti-skeptical strategy and especially the concept of hinges. An influential Wittgenstein-inspired anti-skeptical position is Crispin Wright s rational entitlement strategy ([53], [54]), which can be summarized as follows 4. On Wright s account, hinges such as, There are external objects, Human beings have bodies or The world existed long before my birth are beliefs whose rejection would rationally necessitate extensive reorganisation, or the complete destruction, of what should be considered as empirical evidence and more generally of our epistemic practices. As per Wright, then, each and every one of our ordinary inquiries would then rest on ungrounded presuppositions, hinges ; but still, since the warrant to hold Moore s obvious truisms was acquired in an epistemically responsible way, we could not dismiss them simply because they were groundless as this would lead to complete cognitive paralysis ([53], 191). Following this reading of OC, then, Cartesian skepticism can only show that everyday epistemic practices rest on ungrounded presupposition. But a system of thought, purified of all liability to Moore s obvious truisms of the commonsense, would not be that of a rational agent; thus, we have a default rational basis, an entitlement, to believe in hinges. In this way, Wright argues, we are able to know the denials of skeptical hypotheses; and if we can have some sort of knowledge of hinges such as Human beings have bodies or There are external objects, we are also able to retain confidence in our everyday empirical knowledge claims. A first problem for the entitlement strategy (see [29], [12] and [27]) is that Wright seems to miss a crucial distinction between practical and epistemic rationality. That is, to accept a non-evidentially warranted hinge would be practically rational, as we obviously need to set aside Cartesian skeptical concerns to pursue any kind of inquiry and to achieve cognitive results ([12], 26). But Cartesian skeptical scenarios are not meant to put under discussion the practical rationality of taking for granted Moore s obvious truisms of the commonsense ; rather, they are meant to assess the epistemic rationality of trusting our senses when it is impossible to refute a skeptical scenario such as the BIV one. Thus, even if it would be entirely rational to set aside skeptical concerns whenever we wanted to pursue a given epistemic practice, a Cartesian skeptic can nonetheless argue that the fact that we need true beliefs about the world does not make our acceptance of hinges epistemically rational. Moreover, as has been highlighted by Pritchard ([32], [33]), a crucial

60 58 KRITERION Journal of Philosophy, 2015, 29(2): issue for the entitlement strategy is the very idea of a belief being rationally grounded in something like an entitlement. This is so because to believe a proposition is to believe that proposition to be true; and if this is right, then it is hard to understand how we can have a rational entitlement to take a hinge for granted without having any reason to consider the hinge at issue to be true. Accordingly, if we cannot say, strictly speaking, that we believe in a hinge, for we have no reason to consider it true, then we cannot have knowledge of it either; a mere trusting of or acceptance in the hinge at issue will not suffice. And if this is right, then we cannot have knowledge of hinges, even if we are rationally entitled to take them to be true. 4 Pritchard on the Structure of Reason Wittgenstein s reflections on the structure of reason have influenced a more recent Wittgenstein-inspired anti-skeptical position, namely Pritchard s hinge-commitment strategy (forthcoming). To understand his proposal, recall the following remarks we have already quoted supra: If you are not certain of any fact, you cannot be certain of the meaning of your words either [... ] If you tried to doubt everything you would not get as far as doubting anything. The game of doubting itself presupposes certainty (OC ). The question that we raise and our doubts depend on the fact that some propositions are exempt from doubt, are as it were the hinges on which those turn [....] that is to say, it belongs to the logic of our scientific investigations that certain things are in deed not doubted [...] If I want the door to turn, the hinges must stay put (OC 341-3). As per Pritchard, here Wittgenstein would claim that the same logic of our ways of inquiry presupposes that some propositions are excluded from doubt; and this is not irrational or based on a sort of blind faith, but rather belongs to the way rational inquiries are put forward (see OC 342) 5. As a door needs hinges in order to turn, any rational evaluation would require prior commitment to an unquestionable hinge/set of hinges in order to be possible at all. A consequence of this thought (forthcoming, 3) is that any form of universal doubt such as the Cartesian skeptical one is constitutively im-

61 Nicola Claudio Salvatore: Wittgensteinian Epistemology 59 possible 6 ; there is simply no way to pursue an inquiry in which nothing is taken for granted. In other words, the same generality of the Cartesian skeptical challenge is based on a misleading way of representing the essentially local nature of our enquiries. A proponent of Cartesian skepticism looks for a universal, general evaluation of our beliefs; but crucially there is no such thing as a general evaluation of our beliefs, whether positive (anti-skeptical) or negative (skeptical), for all rational evaluation can take place only in the context of hinges which are themselves immune to rational evaluation. An important consequence of Pritchard s proposal is that it will not affect Closure. Each and every one of our epistemic practices rests on hinges that we accept with a certainty that is the expression of what Pritchard calls über-hinge commitment : an a-rational commitment toward our most basic belief that, as we mentioned above, is not itself opened to rational evaluation and that importantly is not a belief. To understand this point, just recall Pritchard s criticism of Wright s rational entitlement. As we have seen, Wright argues that it would be entirely rational to claim that we know Moore s obvious truisms of the commonsense whenever we are involved in an epistemic practice which is valuable to us; but Pritchard argues that in order to know a proposition we need reasons to believe that proposition to be true. And as, following Wright, we have no reason to consider hinges true but the fact that we need to take them for granted, then we cannot have knowledge of them either. With these considerations in mind, we can come back to Pritchard s über-hinge commitment. As we have seen, this commitment would express a fundamental a-rational relationship toward our most basic certainties, a commitment without which no knowledge is possible. Crucially, our basic certainties are not subject to rational evaluation: for instance, they cannot be confirmed or disconfirmed by evidence and thus they are non-propositional in character (that is to say, they cannot be either true or false). Accordingly, they are not beliefs at all. This can help us retain both the Closure principle and our confidence in our most basic certainties. Recall the reformulation of the Closure principle we have already encountered supra: The Competent Deduction Principle If S knows that p, and S competently deduces from p that q, thereby coming to believe that q on this basis, while retaining her knowledge that p, then S knows that q. The crucial aspect of this principle to note ([33], 14) is that it involves an

62 60 KRITERION Journal of Philosophy, 2015, 29(2): agent forming a belief on the basis of the relevant competent deduction; the idea behind Closure is in fact that an agent can came to acquire new knowledge via competent deduction, where this means that the belief in question is based on that deduction. Accordingly, if we could not rule out a skeptical scenario such as the BIV one, we would be unable to know Moore s obvious truisms of the commonsense such as, Human beings have bodies or There are external objects and thus, given Closure, we would be unable to know anything at all. But our most basic certainties are not beliefs; rather, they are the expression of a-rational, non-propositional commitments. Thus, the skeptic is somewhat right in saying that we do not know Moore s obvious truisms of the commonsense ; but this will not lead to skeptical conclusions, for our hinge commitments are not beliefs so they cannot be objects of knowledge. Therefore, the skeptical challenge is misguided in the first place. A first concern that can be raised against this proposal goes as follows. Recall that, following Pritchard s account, the skeptical challenge is based on a misleading way of representing the nature of our epistemic inquiries; as there is nothing like the kind of general enquiry put forward by a Cartesian skeptic, we should rule out skeptical worries for they are at odds with the ways in which rational inquiries are put forward. However, a skeptic can surely grant that our everyday enquiries are essentially local in nature and that our ordinary knowledge claims are made within a background of hinge-commitments ; but this is just a reflection of what epistemic agents do in normal circumstances, and can at most tell us how our psychology works whenever we are involved in any given epistemic practice. Still, the mere fact that ordinarily we take for granted several hinge commitments does not necessarily exclude as illegitimate the kind of general, theoretical inquiry put forward by a proponent of Cartesian skepticism; for the Cartesian skeptical challenge is first and foremost a philosophical paradox, which cannot be dismissed on the basis of pragmatic reflections about the essentially local nature of our everyday epistemic practices. However, even if we agree with Pritchard that a general evaluation of our beliefs is somewhat impossible and self-refuting there is still another deep concern that the hinge commitment strategy has to face. Recall that following this proposal, all our epistemic practices rest on unsupported commitments. If this approach can help us to block the skeptical challenge it will nonetheless have a cost: under skeptical scrutiny, we will be forced to admit that all our epistemic practices rest on ungrounded

63 Nicola Claudio Salvatore: Wittgensteinian Epistemology 61 presuppositions which are not open to epistemic evaluation of any sort. When skeptical hypotheses are in play, we are then forced to admit that all our knowledge rests on nothing but a-rational presuppositions such as habit, instinct and social or cultural commitments. Accordingly, Pritchard s hinge-commitment strategy leads to a more subtle form of skepticism which undermines the rationality of our ways of inquiry: a conclusion which is no more reassuring than skepticism itself 7. 5 Certainty vs Knowledge Another influential account of Wittgenstein s anti-skeptical strategy is Moyal-Sharrock s non-epistemic reading 8 ; in order to understand this proposal, consider the obvious truisms of the commonsense listed by Moore: There exists at present a living human body, which is my body. This body was born at a certain time in the past, and has existed continuously ever since, though not without undergoing changes; it was, for instance, much smaller when it was born, and for some time afterwards, than it is now. Ever since it was born, it has been either in contact with or not far from the surface of the earth; and, at every moment since it was born, there have also existed many other things, having shape and size in three dimensions [... ] from which it has been at various distances... ([16], 33). What all these statements have in common is that they refer to the empirical world (physical objects, events, interactions) and so they look like empirical propositions. But, argues Moyal-Sharrock, differently from empirical claims they are unquestionable, indubitable and nonhypothetical ([18], 85) statements that cannot be confirmed or falsified by experience; and as Wittgenstein states in his Cambridge Lectures ([50], 16, quoted in [18], 92), a statement which no experience will refute is a rule of grammar : [... ] The proposition describing this world-picture might be a part of a kind of mythology. And their role is like that of rules of a game (OC 95, my italics). When Moore says he knows such and such, he is really enumerating a lot of empirical propositions which we affirm without special testing; propositions, that is, which have a peculiar logic role in the system of our empirical propositions (OC 136).

64 62 KRITERION Journal of Philosophy, 2015, 29(2): Despite their differences, then, for Moyal-Sharrock all hinges share a common feature: namely, they are all rules of grammar which underpin our language-games. This is why, she argues, Wittgenstein considers Moore s knowledge claims in both DCS and PEW as misleading if not completely wrong; for differently from empirical beliefs, hinges cannot be known. To clarify this matter, consider the following entry: And now if I were to say It is my unshakeable conviction that etc, this means in the present case too that I have not consciously arrived at the conviction by following a particular line of thought [my italics], but that it is anchored in all my questions and answers, so anchored that I cannot touch it (OC 103). As per Moyal-Sharrock, this entry highlights the peculiarity of our relationship with hinges. Our taking them for granted is not based on justification or grounds; for instance, I cannot say that I have good grounds for the opinion that cats do not grow on trees or that I had a father and a mother (OC 282). That is, we hold these beliefs unreflectively, for they are not the result of any inquiry and they cannot be supported by any kind of evidence. Still, our lack of grounds for holding hinges does not entail the dramatic conclusions of the Cartesian skeptic, for our relationship with Moore s commonsense certainties is based on training, instinct, repeated exposure ([20], 9): that is, hinges are the result of pre-rational, still perfectly legitimate commitments and are the expression of what Moyal-Sharrock ([18], [19], [20], [21]) calls objective certainty ([18], 15-17). This is a concept that she sees as constitutively different from knowledge; knowledge-claims, in fact, require grounds and/or justifications, are open to doubt and can be verified or disconfirmed by evidence. On the contrary, our confidence in the hinges...lie[s] beyond being justified and unjustified; as it were, as something animal. (OC 359). As per Moyal-Sharrock, our relationship with the hinges is not epistemic or rational at all (hence non-epistemic reading ); following her notion of objective certainty our confidence in the hinges should be seen as kind of doxastic attitude, both as a disposition and an occurrence ([18], 54-56). Quoting Wittgenstein: It is just like directly taking-hold of something, as I take hold of my towel without having doubts (OC 510).

65 Nicola Claudio Salvatore: Wittgensteinian Epistemology 63 And yet this direct taking-hold corresponds to sureness, not to a knowing (OC 511). On Moyal-Sharrock s reading, these remarks suggest that our objective certainty is akin to instinctive or automatic behavior: to a direct taking hold or thought-less grasp ([18], 62). That is to say, this certainty is a disposition of absolute, animal confidence that is not the result of reasoning, observation or research but is rather a basic attitude of unreasoned, unconscious trust that shows itself in our everyday experiences. In other words, our confidence in Moore s obvious truisms of the commonsense such as, There are external objects or Human beings have bodies is not a theoretical or presuppositional certainty but a practical certainty that can express itself only as a way of acting (OC 7, 395); for instance, a hinge such as Human beings have bodies is the disposition of a living creature, which manifests itself in her acting in the certainty of having a body ([18], 67), and manifests itself in her acting embodied (walking, eating, not attempting to walk through walls etc). Following Moyal-Sharrock s account of Wittgenstein s strategy, Cartesian-style skepticism is the result of a Categorial Mistake 9. That is, Cartesian skeptical arguments, even if prima facie compelling, rest on a misleading assumption: the skeptic is simply treating hinges as empirical, propositional knowledge-claims while on the contrary they express a pre-theoretical animal certainty, which is not subject to epistemic evaluation of any sort. Due to this categorial mistake, a proponent of Cartesian Skepticism conflates physical and logical possibility ([18], 170). That is to say, skeptical scenarios such as the BIV one are logically possible but just in the sense that they are conceivable; in other words, we can imagine skeptical scenarios, then run our skeptical arguments and thus conclude that our knowledge is impossible. Still, skeptical hypotheses are nothing but fictional scenarios and once we conflate the logical possibility with the human possibility of being a BIV, then we are making a categorial mistake ([18], ). A consequence of this thought is that Cartesian skeptical scenarios depict a fictional possibility, not a human one; thus, the skeptical challenge is neither a sensible nor legitimate doubt but rather an idle mouthing of words ([18], 174). The mere hypothesis that we might be disembodied brains in the vat has no strength against the objective certainty of hinges such as, There are external objects or Human beings have bodies, just as merely thinking that human beings can fly unaided has no strength against the fact that human beings cannot fly without

66 64 KRITERION Journal of Philosophy, 2015, 29(2): help. Therefore, skeptical beliefs such as I might be a disembodied BIV or I might be the victim of an Evil Deceiver are nothing but beliefbehaviour ([18], 176) and the conclusion we can draw from them, namely that our knowledge is impossible, should be regarded as fiction and not as a possibility: There are contexts then, for the most part: fictional contexts, where the doppelgänger of a universal hinge constitutes a falsifiable proposition. But the negation of a fictional proposition does not entail the negation of any of its doppelgänger. I do not know whether I am a human being pronounced in ordinary circumstances is nonsense. It is not nonsense when pronounced in a fictional context. The problem is that philosophers illegitimately transfer the meaningfulness inherent in the fictional situation to real-life situations ([18], 170, my italics). Following the non-epistemic reading, then, Wittgenstein would dismiss Cartesian-style skepticism as the result of a categorial mistake, based on a confusion between imagined and human/logical possibility. Differently from Wright then, according to Moyal-Sharrock, hinge certainties such as, There are external objects and Human beings have bodies are conceptually, rather than practically, indubitable ([18], 161), whereas the empirical doppelganger of a hinge (i.e. a sentence made up of the same words as a hinge, but which does not function as a hinge) can be doubted. So in ordinary and philosophical contexts hinges can t be doubted; but the same sentence used as an empirical proposition in a sci-fi novel can be. Accordingly, as long as we take skeptical hypotheses as fictional scenarios they make sense but their apparent intelligibility is conflated with human possibility. For instance, the BIV hypothesis is a scenario but is just a fictional one that cannot be applied to our human form of life ; in the world as we know it we cannot even sensibly conceive the existence of bodiless brains connected to supercomputers, or the existence of Evil Deceivers that systematically deceive us and so forth ([18],178). Thus, the strength of Cartesian-style skepticism is, so to speak, only apparent; and once we take skeptical hypotheses as mere philosophical fiction, we should simply dismiss skeptical worries, for a fictional scenario such as the BIV one does not and cannot have any consequences whatsoever on our epistemic practices or, more generally, on our human form of life.

67 Nicola Claudio Salvatore: Wittgensteinian Epistemology 65 This part of the non-epistemic reading seems weak for a number of reasons. If, from one side, Moyal-Sharrock stresses the conceptual, logical indubitability of Moore s truisms, she nonetheless seems to grant that the certainty of hinges stems from their function in a given context, to the extent that they can be sensibly questioned and doubted in fictional scenarios where they can play the role of empirical propositions. But crucially, if hinges are objectively certainty because of their role in our ordinary life, a skeptic can still argue that in the context of philosophical inquiry Moore s commonsense certainties play a role which, similar to the role they play in fictional scenarios, is both at odds with our human form of life and still meaningful and legitimate. Moreover, despite Moyal-Sharrock s insistence on the conceptual, logical indubitability of Moore s truisms of the commonsense, her rendering of Wittgenstein s strategy seems to resemble Wright s proposal ([39],45), thus incurring the objections I have already raised against this reading. As I have argued throughout this work, to simply state that Cartesian skepticism has no consequence on our human form of life sounds like too much of a pragmatist response against the skeptical challenge. This is so because a skeptic can well agree that skeptical hypotheses have no consequence on our everyday practices or that they are just fictional scenarios; also, she can surely grant that Cartesian-style arguments cannot undermine the pre-rational confidence with which we ordinarily take for granted Moore s obvious truisms of the commonsense. But crucially, and as Wittgenstein was well aware, a skeptic can always argue that she is not concerned with practical doubt (OC 19) but with a, so to speak, purely philosophical one. Also and more importantly, even if we agree with Moyal-Sharrock on the nonsensical nature of skeptical doubts, this nonetheless has no strength against Cartesian style skepticism. Recall the feature of Cartesian skeptical arguments: take a skeptical hypothesis SH such as the BIV one and M, a mundane proposition such as, This is a hand. Now, given the Closure principle, the argument goes as follows: (S1) I do not know not-sh (S2) If I do not know not-sh, then I do not know M Therefore (SC) I do not know M In this argument, whether an agent is seriously doubting if she has a body or not is completely irrelevant to the skeptical conclusion I do

68 66 KRITERION Journal of Philosophy, 2015, 29(2): not know M. Also, a proponent of Cartesian-style skepticism can surely grant that we are not BIV, or that we are not constantly deceived by an Evil Genius and so on. Still, the main issue is that we cannot know whether we are victim of a skeptical scenario or not; thus, given Closure, we are unable to know anything at all. Moyal-Sharrock does not explicitly discuss this issue, but her nonepistemic reading so construed seems to leave us with two options, neither of which is particularly appealing. If we stress the non-epistemic nature of hinges while claiming that Cartesian skeptical hypotheses have no strength whatsoever against our knowledge claims, we will be forced to reject a very intuitive principle such as Closure. 10 If, on the other hand, we do not want to reject Closure, it is hard to see how the non-epistemic reading can help us to solve the skeptical problem. For the conclusion we can draw from this proposal is that Cartesian skepticism is unlivable and at odds with our everyday experience; but still, given Closure and the fact that we cannot know the denials of skeptical scenarios, it is impossible to escape skeptical conclusions. Even if more promising than the other OC inspired anti-skeptical proposals I have considered so far, it seems that, nonetheless, the nonepistemic reading cannot represent a satisfactory anti-skeptical strategy. Nevertheless, there are many promising insights we can draw from Moyal-Sharrock s interpretation of Wittgenstein s thought and especially from the analogy between hinges and rules of grammar, which I will consider in the next section. 6 Hinges and Rules of Grammar Very generally, in the second phase of his thought, Wittgenstein calls rules of grammar the conditions, the method necessary for comparing a proposition with reality ([49], 88). Thus, for Wittgenstein, everything that determines the sense of an expression belongs to its grammar, which also specifies the licit combinatorial possibilities of an expression (for instance, which combinations make sense and which don t, which are allowed and which are not allowed, cfr. [13], 146). To understand this point, consider the following statements: i) What is red must be colored ii) Nothing can be red and green all over

69 Nicola Claudio Salvatore: Wittgensteinian Epistemology 67 iii) All bachelors are unmarried iv) A proposition is either true or false v) 12 x 12 = 144 Despite their differences, all these share a number of significant common features. Firstly, they are all normative as they delimit what it makes sense to say, for instance, licensing and prohibiting inferences. Just consider i): if p is called red is correctly characterised as colored, to say that it is red and to deny that it is colored would be a misuse of language, that is, a move excluded from a language-game. Similarly ii), even if it looks as if it is a description of the physics of color, is in fact a rule that we use to exclude the description of an object as being red and green all over. iii) looks like a true statement of fact about bachelors but is, rather, meant to explain the meaning of the word bachelor. iv) looks like a description, a generalization about propositions in the same way that the statement All lions are carnivorous is a generalization about lions. However, things are somewhat different, for we use iv) to define what may be correctly called a proposition in logical reasoning; also, it does not exclude a third possibility but rather excludes as meaningless the phrase a proposition which is neither true nor false 11. Finally 12, central to Wittgenstein s philosophy of mathematics is the view that mathematically necessary truths are not descriptive but normative; for instance, v) licenses and prohibits inferences, in the sense that it licenses transformations of empirical statements and at the same time excludes other inferences as invalid. Following v) 13, we can legitimately transform the statement: There were 12 books each on 12 shelves in the bookshop into: There were 144 books in the bookshop ; also, v) excludes as illegitimate, There were 12 books each on 12 shelves, so there were 1212 books in the bookshop (an inference which is also excluded by the true inequation ). A second feature of Wittgenstein s rules of grammar is that they cannot be confirmed or disconfirmed by reality; rather, they determine what counts as a possible description of reality. That is to say, statements like i) and ii) cannot be confirmed by empirical evidence, but are, rather, presupposed by any language game with color words; also, these grammatical rules cannot possibly be disconfirmed by reality, say by the existence of a colorless red object or of something that is red and green all over. Likewise, we could not verify iii) by, for instance, investigating the marital status of people identified as bachelors, and no married bachelor could possibly disconfirm iii).

70 68 KRITERION Journal of Philosophy, 2015, 29(2): Similarly, even if we do perfectly well speak of half-truths, or rough or approximate truths or of something being partly true or partly false, this does not affect iv) in any way, for the objects of such assertions are not cut to the pattern required for logical inference and thus cannot be considered propositions; therefore, these assertions cannot confirm or disconfirm iv) ([13], 265). Finally, even if we could imagine a different arithmetic in which v) could turn out to be wrong and v*) 12 x 12 = 1212 was correct, this would not disconfirm v), for this v*) would simply not belong to the practice we call arithmetic. A third and important feature of Wittgenstein s rules of grammar is that they are not propositions; namely, they cannot be either true or false for their negation is not false but senseless. Just consider the following sentences: i*) p is red and is not colored ii*) p is red and green all over iii*) Some bachelors are married iv*) a proposition is neither true nor false v*) 12 x 12 = 1212 All these are nothing but nonsensical, even if intelligible, combinations of signs 14. If this is obvious for the putative statements from i*) to iv*), then, as per Wittgenstein, even an equation such as v*) is senseless rather than simply false. As he argued in one of his lectures: The application of a mathematical sentence occurring in our language is not to show us what is true or false but what is sense and what is nonsense. This holds for all mathematicsarithmetic, geometry, etc. For example, there are mathematical propositions about ellipses which show that I cut the elliptical cake in 8 equal parts does not make sense. And there are mathematical propositions about circles which show that it does make sense to say I cut the circular cake in 8 equal parts. The terms sense and nonsense, rather than the terms true and false, bring out the relation of mathematical propositions to non-mathematical propositions (AWL 152) 15. Thus, the difference between rules of grammar and their negations is not similar to the difference between true and false statements, but to the

71 Nicola Claudio Salvatore: Wittgensteinian Epistemology 69 difference between a rule of expression and a use of words or symbols which that rule excludes as nonsensical. 7 Hinges and the Boundaries of Rational Agency To sum up, Wittgenstein s rules of grammar have three features which make them different from empirical beliefs. Firstly, they are not descriptive but normative; secondly, they cannot be confirmed or disconfirmed by reality but, rather, are ways to make sense of reality; finally, they are not propositions as their negations are not false but senseless. With these points in mind, consider the following passages of OC, in which Wittgenstein explicitly compares Moore s obvious truisms to mathematical truths: So is the hypothesis possible, that all the things around us don t exist? Would that not be like the hypothesis of our having miscalculated in all our calculations? (OC 55) I cannot be making a mistake about 12x12 being 144. And now one cannot contrast mathematical certainty with the relative uncertainty of empirical propositions. For the mathematical proposition has been obtained by a series of actions that are in no way different from the actions of the rest of our lives, and are in the same degree liable to forgetfulness, oversight and illusion [... ] The mathematical proposition has, as it were officially, been given the stamp of incontestability. I.e.: Dispute about other things; this is immovable - it is a hinge on which your dispute can turn (OC , my italics). If mathematical truths, qua rules of grammar, license possible ways to describe reality, then to deny or doubt a rule such as v) 12x12=144 does not display factual ignorance but rather the inability to competently engage in the language game called arithmetic. As I have already mentioned throughout this work, for Wittgenstein, the game of doubting itself presupposes certainty (OC 115), that is that something is taken for granted, at least the meaning of words (OC 676). Accordingly, the skeptic s never-ending doubt will deprive her words of their meaning and will, at most, show her inability to engage in the ordinary language-game of asking meaningful questions, as to deny or doubt that i) What is red must be colored and ii) Nothing can be red and

72 70 KRITERION Journal of Philosophy, 2015, 29(2): green all over will display an agent s inability to engage in any sensible language game with color words. This part of Wittgenstein s proposal could be said to resemble Wright s rational entitlement strategy and it incurs similar problems. Recall that, following Wright, we are rationally entitled to take for granted hinges such as Human beings have bodies or There are external objects, for to dismiss them will end up with the impossibility of pursuing any inquiry at all; as practical rational agency is a basic way for us to act, it would be rational to say that we knew hinges, even if in an unwarranted way. As we have extensively seen supra, even if to take hinges for granted would be practically rational, for not to do so would lead to a cognitive paralysis, a skeptic could nonetheless argue that the very fact that we need to act as if hinges are true does not make this acceptance epistemically rational. But following Wittgenstein s reflection on the normative nature of hinges, not to doubt or deny Moore s obvious truisms is not something that we do merely out of practical considerations as in Wright s proposal; rather, it is a constitutive part of the essence of the language-game called epistemic inquiry (OC 370): I want to say: propositions of the form of empirical propositions, and not only propositions of logic, form the foundation of all operating with thoughts (with language) [... ] If I say we assume that the earth has existed for many years past (or something similar), then of course it sounds strange that we should assume such a thing. But in the entire system of our language-games it belongs to the foundations. The assumption, one might say, forms the basis of action, and therefore, naturally, of thought (OC , my italics). As per Wittgenstein, hinges such as There are external objects and Human beings have bodies play a basic, foundational role in our system of beliefs, and to take them for granted belongs to our method of doubt and enquiry (OC 151). In other words, even if they resemble empirical propositions or their origin is empirical, within our practices they are used as rules which enable us to make sense of reality, thus drawing a line between sense and nonsense rather than between truth and falsity. Thus, to doubt or deny Moore s obvious truisms of the commonsense would not only go against our practical rationality, but more crucially would also undermine the very notion of rational enquiry.

73 Nicola Claudio Salvatore: Wittgensteinian Epistemology 71 8 Wittgensteinian Epistemology and Cartesian Skepticism As we have seen, then, for Wittgenstein, Moore s commonsense certainties are a condition of possibility of any meaningful inquiry; as he puts the matter, about certain empirical propositions no doubt can exist if making judgments is to be possible at all (OC 308, my italics). A thought which is stressed in a number of remarks of OC, where Wittgenstein defines hinges as the scaffolding of our thoughts (OC 211), foundation-walls (OC 248), the substratum of all our enquiring and asserting (OC 162), the foundation of all operating with thoughts (401) and fundamental principles of human enquiry (OC 670). To understand a first promising anti-skeptical consequence of this account, recall the putative negation of the rules of grammar we have encountered supra: i*) p is red and is not colored ii*) p is red and green all over iii*) Some bachelors are married iv*) a proposition is neither true nor false v*) = 1212 As we have already seen above, Wittgenstein s rules of grammar are non-propositional in character, thus they cannot be either true or false; accordingly, their negation is not false but senseless, that is, an illicit combination of signs. In a similar fashion, as hinges such as Human beings have bodies or There are external objects are not propositional, for they have a normative rather than a descriptive role, then their putative negation should be dismissed as an illicit (and not only fictional as in Moyal- Sharrock s proposal) combination of signs which is excluded from the practice called rational epistemic inquiry, as the putative statement v*) = 1212 is a move excluded from the language-game called arithmetic. To understand a promising anti-skeptical implication of this point, recall the feature of Cartesian-style arguments: (S1) I do not know not-sh (S2) If I do not know not-sh, then I do not know M

74 72 KRITERION Journal of Philosophy, 2015, 29(2): (SC) I do not know M where not-sh can be a hinge such as Human beings have bodies or There are external objects. This argument seems compelling as long as we take hinges as propositional beliefs which can be either confirmed by evidence or legitimately doubted once we run skeptical arguments. But even if they resemble empirical contingent propositions, hinges are nonpropositional rules of grammar which enable us to make sense of reality. Accordingly, skeptical hypotheses such as I might be a disembodied BIV should not be regarded as sensible philosophical challenges but rather as nonsensical, even if prima facie meaningful, combinations of signs. Another promising consequence of a non-propositional account so construed is that, different from Moyal-Sharrock s reading of OC, it will not affect the Closure principle and at the same time will not lead to skeptical conclusions. Recall that following the non-epistemic reading, the certainty of hinges is a pre-rational, animal commitment which is not subject to epistemic evaluation of any sort. Accordingly, following this account, we will have either to reject Closure or, with this principle still in play, to agree with the skeptic that our knowledge is impossible. Consider the formulation of Closure proposed by Williamson ([44]) and Hawthorne ([14]) which we have encountered throughout this work: The Competent Deduction principle If S knows that p, and S competently deduces from p that q, thereby coming to believe that q on this basis while retaining her knowledge that p, then S knows that q. The idea behind this version of Closure is in fact that an agent can come to acquire new knowledge via competent deduction where this means that the belief in question is based on that deduction. Accordingly, if we cannot rule out a skeptical scenario such as the BIV one, given Closure we would still be unable to know anything at all. The non-propositional nature of Wittgenstein s account of hinges can help us to positively address this issue. As we have seen while presenting Pritchard s hinge-commitment strategy, the crucial aspect to note about Closure is that it involves an agent forming a belief on the basis of the relevant competent deduction. But crucially, hinges are not the expressions of a propositional attitude such as a belief in; rather, they are the expression of non-propositional rules. 16 Accordingly the negations of hinges, that is, skeptical hypotheses such as I might be a disembodied BIV or I might be deceived by an

75 Nicola Claudio Salvatore: Wittgensteinian Epistemology 73 Evil Demon are not beliefs either; rather, they are just nonsensical combinations of signs, from which no valid inference or deduction (e.g. If I do not know not-sh, then I do not know M ) can be made. That is to say, if skeptical hypotheses are not propositional beliefs but rather, senseless negations of non propositional rules, then from the fact that we don t know whether we are victims of a skeptical scenario we cannot infer or deduce that we don t know everyday empirical propositions ; we are thus in a position to retain Closure (which can be applied only to propositional beliefs, and not to nonsensical negations of non propositional rules) and our confidence in our everyday knowledge claims. 9 Concluding Remarks In this paper, I have argued that following the analogy between hinges and rules of grammar, we should be able to get rid of skeptical hypotheses. This is so because given the non propositional, normative nature of hinges such as There are external objects or Human beings have bodies their skeptical negations are not propositional beliefs but rather, nonsensical, even if apparently meaningful, combinations of signs from which no valid inference or deduction can be made 17. Notes 1 See [35]. 2 This is essentially the formulation of the Closure principle defended by [44], 117, and [14], While writing OC Wittgenstein was also heavily influenced by Henry Newman s lectures on religious beliefs (see [24], [25]). For a more detailed analysis of the relationship between Newman s and Wittgenstein s anti-skeptical strategies, see [28]. 4 Wright s proposal is informed by his diagnosis of Moore s Proof ([52]), from which has originated a huge debate that would be impossible to summarize here. For a critical analysis of Wright s treatment of PEW see [36], [37], [38], [7] and [2], [3]. 5 Cfr OC 342: [... ] it belongs to the logic of our scientific investigations that certain things are indeed not doubted. 6 See OC 450: A doubt that doubted everything would not be a doubt. 7 Pritchard has explicitly addressed this issue, stating that even if his proposal blocked the skeptical challenge it would nonetheless lead to what he names epistemic angst or, more recently, epistemic vertigo. See [31]. Moreover, it should be noted that Pritchard s reflections on hinges are only a part of a more complex anti-skeptical framework.; the other part is called epistemological disjunctivism ([34]). As in this work I am focusing only on Wittgenstein inspired

76 74 KRITERION Journal of Philosophy, 2015, 29(2): anti-skeptical proposals, to present and discuss the merits of Pritchard s epistemological disjunctivism would go beyond the scope of this essay and is thus not a task I shall set myself here. 8 It is worth noting that Pritchard ([30]) calls his own anti skeptical strategy nonepistemic (268) and defines Moyal Sharrock s proposal non-propositional (265). Still, Moyal Sharrock stresses the non epistemic, pre rational nature of hinges while as we have seen above Pritchard focuses on the non propositional nature of what he names uber hinge commitments ; hence I preferred to call the former strategy non epistemic. 9 See OC 308: Knowledge and Certainty belong to different categories. They are not two mental states like; say surmising and being sure. (Here, I assume that it is meaningful for me to say, I know what (e.g.) the word doubt means and that this sentence indicates that the word doubt has a logical role.) What interests us now is not being sure but knowledge. That is, we are interested in the fact that about certain empirical propositions no doubt can exist if making judgments is to be possible at all. Or again: I am inclined to believe that not everything that has the form of an empirical proposition is one. 10 This line has been most notably proposed by [8], [9], [10], [11]) and [26]. 11 According to the proponents of many-valued logic such as [43], statements of the form a proposition which is neither true nor false are borderline cases, whose truth value lies between 0 (full falsehood) and 1(full truth); thus, they would not be mere senseless combinations of signs as in Wittgenstein s account. Even though this approach has been extremely useful to deal with a number of philosophical issues such as the vagueness problems, this view is still far from uncontroversial and has originated a huge debate that would be impossible to summarize here. For an up to date discussion on multi-valued logic and the vagueness problem, see [42]. 12 In the following, I will just sketch some uncontroversial aspects of Wittgenstein s philosophy of mathematics, in order to cast more light on his conception of rules of grammar. A detailed reconstruction of Wittgenstein s views on the matter and of the debate they originated would fall beyond the scope of this work. 13 This is a slightly modified example used by [13], It is worth noting that Wittgenstein considers senseless every combination of signs excluded from a rule of grammar. This is so because as grammatical rules are ways to make sense of reality, their correctness is antecedent to questions of truth of falsity and so they lack a truth-value. Accordingly, their putative negations lack truth-value as well; thus, they cannot be considered strictly speaking false but senseless, that is illicit, combinations of signs. 15 It could be argued that there are mathematical and geometrical discoveries, as in the case of Non-Euclidean geometry. Still, as per Wittgenstein, these are different from empirical discoveries for they do not tell us anything about reality but are, rather, different techniques to describe reality. See infra. 16 For a similar account of hinges and their anti skeptical significance, see [4], [5]. Roughly, as per Coliva hinges, even if propositional, have a normative role, and their acceptance is a condition of possibility of any rational enquiry. A first difference between this account and the one I m defending here goes as follows; according to Coliva, hinges are propositional (albeit non bipolar) ; on the contrary, I claim that they are non propositional, hence their putative negations (such

77 Nicola Claudio Salvatore: Wittgensteinian Epistemology 75 as skeptical hypotheses)are senseless and excluded from our epistemic practices. Moreover, and more importantly, Coliva proposes a limitation of the Closure principle ([5], 86; a similar view is defended in [1]), which stems from her views on warrant and epistemic justification that will be impossible to summarize here. However, following my account of hinges there is no need to defend a limitation of Closure; this is because if skeptical hypotheses SH such as I might be a BIV or I might be deceived by an Evil Deceiver are senseless combinations of signs, so are their putative negations not-sh; then from the fact that we don t know whether we are victims of a skeptical scenario ( I do not know not-sh where both SH and its negation are illicit combination of signs )we cannot infer or deduce that we do not know our everyday propositions M even with a strong version of Closure in play. On Coliva s reading of OC and its anti-skeptical implications, see [21] and [31]. 17 For a general introduction to OC see also [23] and [41]. Other influential OC inspired anti-skeptical strategies are [6] and [45], [46], [47]. For a critical evaluation of Conant s and Williams proposals, see [39], [40]. Nicola Claudio Salvatore Universitade Estadual de Campinas UNICAMP Instituto de Filosofia e Ciências Humanas R. Corá Coralina 100 Cidade Universitária Campinas SP, , Brazil <n162970@dac.unicamp.br> References [1] Avnur Y. (2011) An Old Problem for New Rationalism, Synthese, 183/2, [2] Coliva, A. (2009a), Moore s Proof and Martin Davies epistemic projects, Australasian Journal of Philosophy. [3] Coliva, A (2009b), Moore s Proof, liberals and conservatives. Is there a third way? in A. Coliva (ed.) Mind, Meaning and Knowledge. Themes from the Philosophy of Crispin Wright, OUP [4] Coliva, A. (2010). Moore and Wittgenstein. Scepticism, Certainty and Common Sense. Palgrave MacMillan

78 76 KRITERION Journal of Philosophy, 2015, 29(2): [5] Coliva, A. (2015). Extended Rationality. A Hinge Epistemology. Palgrave MacMillan [6] Conant, J. (1998), Wittgenstein on meaning and use, Philosophical Investigations [7] Davies, M. (2004), Epistemic entitlement, warrant transmission and easy knowledge. Aristotelian Society Supplementary Volume, 78:1, July 2004, [8] Dretske, F. (1970), Epistemic Operators. Journal of Philosophy, vol. 67,1970. [9] Dretske, F. (1971), Conclusive Reasons. Australasian Journal of Philosophy, 49, [10] Dretske, F. (2005a), The Case against Closure, in Contemporary debates in Epistemology, eds. M. Steup and E. Sosa, [11] Dretske, F. (2005b), Reply to Hawthorne, in Contemporary debates in Epistemology, M. Steup and E. Sosa (eds.), [12] Jenkins, C. (2007), Entitlement and Rationality, Synthese 157, [13] Hacker P.M.S. and Baker, G.P (1985), Wittgenstein: Rules, Grammar, and Necessity, Volume 2 of an analytical commentary on the Philosophical Investigations, Oxford, UK and Cambridge, Mass. USA: Blackwell. [14] Hawthorne, J. (2005), The Case for Closure, in Contemporary debates in Epistemology, M. Steup and E. Sosa (eds.), [15] Minar, E. (2005), On Wittgenstein s Response to Scepticism: The Opening of On Certainty, in D. Moyal-Sharrock and W.H. Brenner (eds.), Readings of Wittgenstein s On Certainty, London: Palgrave, [16] Moore, G.E. (1925), A defense of Common Sense, in Contemporary British Philosophers, 1925, reprinted in G. E. Moore, Philosophical Papers, London: Collier Books, [17] Moore, G.E. (1939), Proof of an external world, Proceedings of the British academy, reprinted in Philosophical Papers.

79 Nicola Claudio Salvatore: Wittgensteinian Epistemology 77 [18] Moyal-Sharrock, D. (2004), Understanding Wittgenstein s On Certainty, London: Palgrave Macmillan. [19] Moyal-Sharrock, D. and Brenner, W.H. (2005), Readings of Wittgenstein s On Certainty, London: Palgrave. [20] Moyal-Sharrock, D. (2010), Hinge Certainty, draft. [21] Moyal-Sharrock, D. (2013), On Coliva s Judgmental Hinges, Philosophia, Volume 41, Number 1, pp [22] McGinn, M. (1989), Sense and Certainty: A Dissolution of Scepticism, Oxford: Blackwell. [23] Morawetz, T. (1978), Wittgenstein & Knowledge: The Importance of On Certainty, Cambridge, MA: Harvester Press. [24] Newman, J.H. (1844), Sermons, Chiefly on the Theory of Religious Belief, Preached before the University of Oxford, London. [25] Newman, J.H. ( ), An essay in Aid of a Grammar of Assent, ed. I.T. Kerr, Oxford: Oxford University Press. [26] Nozick, R. (1981), Philosophical Explanations. Oxford: Oxford University Press. [27] Pedersen, N. (2009) Entitlement, Value and Rationality, Synthese 171 (3) [28] Pritchard, D. H. (2000), Is God Exists a Hinge Proposition of Religious Belief?, International Journal for Philosophy of Religion 47, [29] Pritchard, D. H. (2005), Wittgenstein s On Certainty and contemporary anti-skepticism, in Readings of Wittgenstein s On Certainty, D. Moyal-Sharrock and W.H. Brenner (eds.), London: Palgrave, [30] Pritchard, D. H.(2012) Wittgenstein and the Groundlessness of Our Believing. Synthese, 189, [31] Pritchard, D. H. and Boult, C. (2013), Wittgensteinian Anti- Scepticism and Epistemic Vertigo, Philosophia, Volume 41, Number 1, pp

80 78 KRITERION Journal of Philosophy, 2015, 29(2): [32] Pritchard, D. H. (2014) Entitlement and the Groundlessness of Our Believing, in Contemporary Perspectives on Scepticism and Perceptual Justification, D. Dodd & E. Zardini (eds.), Oxford UP. [33] Pritchard, D. H (forthcoming a), Wittgenstein on Hinges and Radical Scepticism in On Certainty, Blackwell Companion to Wittgenstein, H. -J. Glock & J. Hyman (eds.), Blackwell. [34] Pritchard, D.H. (forthcoming b) Epistemic Angst. Radical Scepticism and the Groundlessness of Our Believing. Princeton University Press. [35] Putnam, H. (1981), Reason, Truth and History. Cambridge: Cambridge University Press. [36] Pryor, J. (2000), The Skeptic and The Dogmatist, Noûs 34, [37] Pryor, J. (2004), What s wrong with Moore s Argument? Philosophical Issues, 14. [38] Pryor, J. (2012), When warrants transmits in Wittgenstein, Epistemology and Mind: Themes from the Philosophy of Crispin Wright, Annalisa Coliva (ed.), OUP, [39] Salvatore, N.C. (2013), Skepticism, Rules and Grammar, Polish Journal of Philosophy, Vol. VII, No. 1, [40] Salvatore, N.C. (forthcoming), Skepticism and Nonsense, Southwest Philosophical Studies [41] Stroll, A. (1994), Moore and Wittgenstein on Certainty, Oxford: Oxford University Press. [42] Sorensen, R. (Winter 2013 Edition), Vagueness, The Stanford Encyclopedia of Philosophy Edward N. Zalta (ed.), URL [43] Weber, Z. and Colyvan, M., (2010), A topological sorites, The Journal of Philosophy, 107, [44] Williamson, T. (2000), Knowledge and Its Limits. Oxford: Oxford University Press.

81 Nicola Claudio Salvatore: Wittgensteinian Epistemology 79 [45] Williams, M. (2004a), Wittgenstein s refutation of idealism, in Wittgenstein and Skepticism, D. McManus (ed.), London, New York: Routledge, [46] Williams, M. (2004b), Wittgenstein, truth and certainty, in Wittgenstein s lasting significance, M. Kolbel, B. Weiss (eds.), London: Routledge. [47] Williams, M. (2005), Why Wittgenstein isn t a foundationalist, in Readings of Wittgenstein s On Certainty, D. Moyal-Sharrock and W.H. Brenner (eds.), [48] Wittgenstein, L. (1969), On Certainty, ed. G. E. M. Anscombe and G. H. von Wright, tr.d. Paul and G. E. M. Anscombe, Oxford: Blackwell. [49] Wittgenstein, L. (1974), Philosophical Grammar, ed. R. Rhees, tr. A. J. P. Kenny, Blackwell, Oxford. [50] Wittgenstein, L. (1979), Wittgenstein s Lectures, Cambridge , from the Notes of Alice Ambrose and Margaret Mac- Donald, ed. Alice Ambrose, Oxford: Blackwell. [51] Wittgenstein, L. (2009), Philosophical Investigations, revised 4th edn edited by P. M. S. Hacker and Joachim Schulte, tr. G. E. M. Anscombe, P. M. S. Hacker and Joachim Schulte, Oxford: Wiley- Blackwell. [52] Wright, C. (1985), Facts and Certainty, Proceedings of the British Academy 71, [53] Wright, C (2004a), Warrant for nothing (and foundation for free)? Aristotelian society Supplement, 78/1, , [54] Wright, C. (2004b), Wittgensteinian Certainties in Wittgenstein and Skepticism, D. McManus (ed.),

82

83 Why Sufficientarianism is not Indifferent to Taxation Philipp Kanschik Abstract The indifference objection is one of the most powerful objections to sufficientarianism. Critics argue that sufficientarianism is objectionably indifferent to the distribution of benefits and burdens. This article focuses on the criticism of the latter, particularly the claim that sufficientarianism is indifferent to taxation. Contrary to this allegation, it is argued that sufficientarianism warrants progressive taxation, the reason being that even those who are sufficiently well off face the risk of being pushed below sufficiency. This risk decreases the better off someone is as it is easier for those who are better off to deal with sufficiency-threatening circumstances. It is argued that the risk of insufficiency understood as a function of distance to the threshold justifies progressive taxation. The proposed line of reasoning corresponds to the sufficientarian belief that there should be no redistribution between the rich, as the differences in risk of insufficiency eventually become marginal among those who are very well off. Moreover, the proposed rationale for progressive taxation does not depend on prioritarian or egalitarian reasoning. Rather, it transpires that sufficientarianism is well-suited to justify the progressive redistributive system of the modern welfare state. Keywords: sufficientarianism, distributive justice, indifference objection, ethics of taxation The doctrine of sufficientarianism has recently gained some momentum in distributive justice as a rival to prioritarianism and egalitarianism. However, the indifference objection arguably remains one of the most powerful objections to the doctrine. It alleges that sufficientarianism is objectionably indifferent to distributions of benefits and burdens once everyone has secured enough. Specifically, critics maintain that sufficientarians are indifferent between progressive and regressive tax systems. Such indifference runs counter to intuitions toward a progressive tax Kriterion Journal of Philosophy, 2015, 29(2): c 2015 The author