The Epistemological Functions of Metaphysics

Alexei Procyshyn The New School For Social Research The Epistemological Functions of Metaphysics Presented at the 16th Annual Conference of Philosophy Villanova University, April 8th 2011 About one hundred pages into Guerrilla Metaphysics, Graham Harman issues a challenge to his readers, concerning the cogency of object oriented philosophy: “Let anyone who does not agree with the strategies of object oriented philosophy specify clearly which of its initial steps is invalid.”1 Much of what I plan to say today stems from my acceptance of Harman’s challenge and my attempt to formulate a cogent argument as to why I find his approach particularly unsatisfying. By way of a conclusion, I’ll even offer what strikes me as a knockdown argument against his philosophy. But ‘refutation,’ should I turn out to have one, will not have been my primary focus. It simply follows from a more general concern for the role metaphysics plays within our broader intellectual pursuits. For the role we assign to metaphysics will determine the criteria we use to evaluate the validity of a particular metaphysical strategy. So instead of the standard ‘what?’ questions metaphysicians tend to pose, I want to ask, ‘why metaphysics?’ or, perhaps better yet, why is metaphysics an intelligible undertaking? Or again, to broach the theme of this conference, why does metaphysics seem to be a pressing concern today, why has it returned? A respectable metaphysician should, in general, have a reasonable answer to these questions, but here in particular it will set the stakes for Harman’s challenge. Here is my answer: we appeal to metaphysics (and ontology in particular) to delineate the specific commitments we take on when we engage in certain practices, and these commitments, appropriately developed or reconstructed, serve to ground our knowing and doing. Metaphysics, on this view, grounds our (epistemic) practices2 by making explicit the commitments we must accept and thereby defeats skeptical worries or alternative accounts of their intelligibility. Ontology, in short, furnishes us with the conditions of intelligibility for some set of practices or a region of experience. In this sense, metaphysics never really left the stage. 1. Shallowness, Inference & The Parameters of Metaphysics The virtue of this characterization lies in showing how a given metaphysical enterprise can succeed or fail. For, however elegant a metaphysical system might be, in the absence of clear evaluative criteria it remains meaningless. Metaphysics without the possibility of truth or falsity is like colourless green dreams that sleep furiously, or present kings of France who are bald. So let me begin by offering a criterion for evaluating metaphysics, which I will term shallowness.3 Shallowness identifies a pervasive defect in our reasoning about putatively metaphysical features that follows from a failure to recognize that the inferential structure underwriting metaphysics cannot be strictly logical. Shallow metaphysics, in other words, fails to identify genuine commitments and therefore fails to ground any practice. At its most basic, then, the charge of shallowness results from a purely verbal or terminological approach to metaphysical problems. This kind of objection draws attention to the parameters of a given problem space, and shows that the various positions being taken hinge on some kind of legitimate conceptual relativity; in effect, shallowness points out that the very problem space has been poorly delineated. To illustrate, consider what might happen when I hold up my hand and ask, ‘How many fingers?’ A natural response would be to say that I’m holding up five; another, four and a thumb. Now imagine that four-finger ontologists were 1 Alexei Procyshyn The New School For Social Research debating five-finger ontologists. We would, I think, quickly realize that what is at stake is how we define ‘finger’ rather than some metaphysical entity, like ‘fingerness’ or the ‘being of fingers.’ The whole debate appears to rest on a basic mistake, since all the responses to the question concerning the meaning of ‘finger’ accept the same basic features of the world (the physical structure of hands), but parse them differently. What is at stake then are not the facts or the structure of the world that grounds them, but simply how we describe them and apply these descriptions. Shallowness is thus what we get when we try to derive a truth about the world from the meaning of concepts. Not only is there something inherently backwards about that, but also notice that this backwardness affects our inferences. For example, we might be tempted to claim that the assertion, ‘there is at least one number’ follows from the claim that the number of fingers on my hand is finite. Now, the existence of abstract entities like numbers is not a shallow matter – there is a serious ontological commitment there. Moreover, the argument is valid, since the conclusion follows from the premise. A commitment to at least one number analytically follows from ‘there are a finite number of fingers on my hand.’ This argument even appears to be sound, since we can establish the truth of the ground by simply counting our fingers. What makes this argument shallow is that the analytic relation between abstract entity and finite number of fingers is completely trivial: the conclusion is entailed by the premise and the premise is justified by an act of counting. But all this begs the question. All we have done is draw out the conclusion about the existence of at least one number from a feature of a conceptual framework that is already committed to them. So our conclusion turns out to be a point about language or even the nature of representation more broadly,4 but not about the nuts and bolts of a world in which numbers exist. We can see, then, how shallowness might condemn a metaphysical project. For, if an honest metaphysician relies solely on a given terminology, her work becomes little more than a formal game. Worse: her method amounts to a mere teasing out the implications of her concepts. The problem, of course, is that these concepts might not pick out anything. Statements containing them, such as ‘the present king of France is bald’ or ‘colourless green dreams sleep furiously,’ may appear coherent but ultimately lack any truth-value. Or they may have a sense (an intension), but prescribe non-existent entities and hence cannot be true. Shallowness thus offers a litmus test for the basic make-up of metaphysics. For it emphasizes that metaphysical claims need to have truth-values to avoid being shallow, and hence theories need to specify their truth conditions. And, as we saw in the previous examples, these conditions cannot simply be analytic, conceptual, or logical. There is a further consequence to these considerations. I have already mentioned that validity and soundness may not be the appropriate evaluative criteria. This implies a deeper problem about the very structure of metaphysical reasoning. Consider again the example concerning finger ontology: the problem was that the number of fingers on my hand depended upon how we defined ‘finger.’ This means that the claims of four- and five-finger ontologists are true relative to the relevant definition. Hence, we can reformulate the dispute as involving two distinct conditional claims. Four-finger ontologists advance something like, 1. If ‘finger’ means x, then there are four of them on my hand, while proponents of a five-finger framework are committed to, 2 Alexei Procyshyn The New School For Social Research 2. If ‘finger’ means y, then there are five of them on my hand. The verbal nature of the problem now comes into full view, since both conditional claims seem acceptable. On intuitive grounds, there simply is no substantive disagreement: relative to the respective definitions of ‘finger,’ both claims appear to be true. Our intuitive acceptance, however, gives shallowness its teeth, insofar as it rules out any appeal to material implication. We can see why this is so by inspecting the truth-functional characterization of material implication. Thus this truth-table: A T T F F B T F T F A⊃B T F T T ~A ⊃ B T T T F A ⊃ ~B F T T T The paradox may already be clear, but let me state it explicitly: the falsity of the antecedent A (that is, our definition of ‘finger,’ whose truth our toy ontologists are debating) is sufficient for the truth of A ⊃ B.5 This is particularly startling because it now seems that the very debates of finger-ontology have no bearing on the truth of the conditional claim each ontologist advocates. Where the proposed definition fails, the conditional is true, and where the consequent is true the definition is irrelevant.6 But this is just another way of saying that if finger ontology advances material conditionals like (1) and (2), this ontology is inherently shallow, for the definition of the concept ‘finger’ does not play much of a role. Relying on material conditionals thus leads to bad metaphysics, since these conditionals fail to establish an adequate relation between antecedent and consequent. This is not to say we cannot use conditionals: we can and do, all the time; it’s just that they are not strictly logical. The threat of shallowness demonstrates why our traditional notions of validity (even soundness) are insufficient for evaluating metaphysical reasoning, and shows how we can go wrong in our metaphysics. First and foremost, metaphysics cannot be only about concepts. For purely conceptual considerations merely move us from the meaning of a concept to the potential consequences of its possible application, but cannot generate actual applicability. That’s why beginning from, or simply generating new concepts is backwards. Worse, meaningful concepts do not necessarily generate meaningful sentences, and they certainly cannot establish the existence of an entity. Again, it is entirely possible that a given concept plays absolutely no role in our actual cognitive endeavours, or plays a role but fails to denote anything. A case in point would be ‘sympathetic magic,’ which no longer plays a role in our scientific practice, but failed to pick out any actual entities even when it did. 2. Agrippan Skepticism, Ontological Response Shallowness thus shows us what good metaphysics must look like, for it underscores that ontological grounds need to be extrapolated from actual practices in a manner that does not rely solely on the features of the concepts employed. On pain of shallowness, there can be no metaphysics for metaphysics’ sake. But this does not yet answer the ‘why’ question I originally posed; it just specifies what we need to avoid. To get a feel for the necessity of metaphysical inquiry, for why it must be connected to our actual practices, we need to consider the problems 3 Alexei Procyshyn The New School For Social Research facing these practices themselves. How, in other words, do we know when we get something right? In the absence of ontology, this kind of question generalizes into radical skepticism, relativism, or some deflationary stance like quietism or fideism. Ontology offers us an alternative. This, I take it, is the epistemological function of metaphysics. Now, when we combine the anti-skeptical function of grounding practices with an awareness of the problems associated with shallowness, we generate transcendental argumentation. This strategy combines considerations of concrete epistemic practices with a non-formal extrapolation of ontological commitments in its familiar two-step, inferential pattern. This argument form circumvents skepticism by ensuring that the move from antecedent practice to consequent ontological commitment is deep. In fact, I shall argue later that the relation is deep precisely because it specifies what is ‘indispensable’ about a given practice. For now, however, let me establish the convergence of epistemological function with our concerns for shallowness. As both Paul Franks and Michael Williams have shown,7 transcendental arguments respond to the Agrippan trilemma (see AN 17f and UD 60-67). The trilemma consists in showing that any particular claim is vulnerable to skepticism, because our course for establishing it necessarily involves some form of justificatory failure, which undermines the argument as a whole. Once generalized, the trilemma appears to threaten the possibility of justification tout court. When confronted by a skeptical worry, so the Agrippan contends, there seem to be only three response possibilities: (1) Refuse to respond, i.e. make an undefended assumption; (2) Repeat a claim made earlier in the argument, i.e. reason in a circle; (3) Keep trying to think of something new to say, i.e. embark on an infinite regress. Since none of these responses is satisfactory, the trilemma represents a general strategy capable of undermining almost any claim. It targets one’s warrant for asserting a claim and shows that no legitimate ground for asserting it can be produced; it dooms us to groundlessness, vicious circularity, or infinite regress. As Franks notes, the “Agrippan trilemma […] can arise whenever what is at stake is the nature of reasons” (AN 18) – and this includes metaphysics. The key supposition here is that justifications must terminate, or be grounded by something that necessarily exceeds our epistemic relationships. We can thus isolate two connected demands, which Franks explains as follows: The problem [of the Agrippan trilemma] arises from the combination of two demands. First, there is the Monistic Demand. This is the demand that every genuine grounding participate in a single systematic unity of grounds, terminating in a single absolute ground. […] Second, there is the Dualistic Demand that physical grounding and metaphysical grounding be kept rigorously separate. […] The two demands can be seen to be in apparent conflict once the following observation […] is added: physical explanations do not and indeed cannot terminate in an absolute ground. (AN 20) The tension here is characteristic of metaphysics in general. For in teasing out a non-physical, non-causal, non-epistemic – and, as we have already seen, non-conceptual, non-analytic – relationship between a claim to know and its ground, ontology accepts and responds to the Agrippan challenge by seeking to satisfy Franks’ monistic and dualistic demands. Notice, however, that in doing so, ontology must commit to some kind of two-step reasoning structure. 4 Alexei Procyshyn The New School For Social Research There is simply no other way of making good on metaphysics’ anti-skeptical promise. But this also means that one cannot neatly separate ontology from epistemology, and questions concerning the one also affect the other. Slightly rephrased, without a preliminary skeptical concern about a human practice, we cannot motivate ontology, and without ontology there is no secure set of grounds and grounding relations for knowledge.8 The absence of a legitimate 1st order concern directly leads to shallowness – to an empty thought spinning frictionlessly about itself in the void. All this helps us answer our ‘why?’ question concerning metaphysics, while motivating a two-step form of inquiry whereby we delineate the grounds and relations underwriting our practices in order to circumvent skepticism (and its ultimate consequences: quietism, fideism and nihilism). That is, we engage in metaphysics in order to develop a non-empirical system of grounds that vouchsafes knowledge and helps us determine the success-conditions for our endeavours. What remains to be shown is how responding to the dualistic and monistic demands operates relative to the problems of shallowness. 3. Transcendental Arguments Synthesize Conditionals and Demands Productive metaphysics must embrace some kind of two-step strategy for metaphysical reasoning that is not simply logical. This has important implications for the kinds of conditionals we employ in our arguments. For we now need a thickly specified relation between antecedent practice and consequent ontological commitments. And, as I intimated already, the longstanding candidate for this relation has been indispensability. Before addressing what indispensability means and how it works in transcendental argumentation, let us synthesize our results so far. Combining our concern for shallowness with metaphysics’ epistemological functions, we have seen that,     The relation permitting us to step from antecedent to consequent cannot be analytic, or purely syntactic (e.g. logical, truth-functional), nor can it be causal. Our antecedent object cannot be verbal or terminological in nature, but must in fact be some kind of successful practice. The consequent (or ontological commitment) must be distinct in kind from its extrapolated base. Any conditional whose consequent entails the falsity of its antecedent claim generates incoherence. We can see how these features respond to the Agrippan trilemma. The first three ensure that an explanation can terminate without circularity, regress, or dogmatism, while the last one articulates a check or constraint on what kinds of grounds can actually be asserted. All that remains is to specify a non-analytic and non-causal relationship between a practice and its metaphysical grounds. Such a relation defines transcendental argumentation. We already know that transcendental arguments are anti-skeptical in nature, and that they extrapolate from first person experience to make claims about the grounds of experience as such (i.e. conditions of possibility). As Robert Stern explains in his book on the subject, we can “broadly characterize transcendental arguments as involving the claim of the form ‘X is a necessary condition for the possibility of experience, language, thought, etc.’ where the rationes cognescendi of this claim is 5 Alexei Procyshyn The New School For Social Research non-empirical, and the rationes essendi is not that it is analytically true or true by virtue of the laws of nature.”9 This characterization possesses all the features we have identified above: it clearly involves a two-step strategy where we move from an antecedent, non-analytic feature of experience to an ontological entity, while stipulating that this ‘inference’ cannot be shallow or trivial. What I now want to show is that antecedent practice is non-analytically and non-causally related to its consequent grounds via indispensability. At any rate, this has been the go-to relation in the history of philosophy: Descartes, Leibniz, Kant, and Quine, just to name a few prominent examples, have all deployed it. We can explain the notion of indispensability by way of another example.10 Consider the relationship between a true sentence S, and the language in which it is uttered. In an interesting way, the truth of S hinges on the existence of the language, since it couldn’t be true unless it was asserted, and one cannot assert a sentence without a language. Analogously, one cannot truly assert that there is no language, since the very assertion hinges on there being a language in which to utter it and hence contradicts itself (this is the basic notion of performative contradiction). There is then a necessary relationship between a specifically human contribution to experience – uttering a sentence – and the grounds that make such an experience possible (i.e. a language). The indispensable feature of asserting a truth is thus language and so this language’s indispensability generates an ontological commitment. The distinguishing feature of transcendental arguments, which separates them from deductively valid ones, is the manner in which they move from the indispensability of some feature of experience to the ontological commitments which function as the preconditions for our performances. Their two-step structure consists in establishing that a specific feature of an activity is indispensable; second, they show how this indispensable feature commits us to some ontological entity or framework. W.V.O. Quine gave this kind of reasoning the most laconic formulation, when he quipped that “To be is to be the value of a bound variable.”11 His formulation of our indispensability argument has the following form: (1) α is indispensable to our best (epistemic) practices or theories. (2) The structure of α commits us to χ. (3) Therefore we should accept χ as the ground for our best (epistemic) practices or theories. For Quine, of course, indispensability is bound up with reference (i.e. existentially quantified variables with putatively non-empty extension). What fixes reference, according to him, is the system of suppositions implicit in scientific practices12 – this is what Franks understood as an Absolute ground. What indispensability does, then, is allow us to extrapolate the implicit commitments of our practices. With more time, I could show how this strategy has been implemented in the history of philosophy. Let me just point here, by way of its clearest illustration, to Kant’s transcendental deduction, which comprises two steps: a metaphysical deduction, and a transcendental one. The metaphysical deduction establishes the indispensability of pure categories for the acts of judgment, while the transcendental deduction proper shows that these categories commit us to the transcendental unity of apperception, which systematically grounds the cognitive performances comprising the threefold syntheses of apprehension, imagination, and comprehension in a concept. Descartes’ proof for his own existence operates in a similar way: 6 Alexei Procyshyn The New School For Social Research the indispensable condition for the truth of ‘I think’ is an I to assert it; hence we are committed to the ontological structure of this ‘I.’ 4. Metaphysical Fallacies I promised to show why I think Harman’s philosophy fails. To do so, let me first consolidate my foregoing argument: considered in its epistemological functions, a non-shallow metaphysics offers grounds for our best practices. So understood, our reasoning about metaphysical issues must employ a two-step pattern of reasoning. Although this pattern superficially resembles the way we operate with material conditionals and modus ponens, the relationship between antecedent practice and consequent commitments must be non-logical, non-causal, and nonconceptual. And the prime candidate for such a connection is indispensability. I further argued that ‘indispensability’ must be indexed to a first-person perspective and a set of successful (epistemic) practices.13 When these restrictions on indispensability fail, say in the case of sympathetic magic, the notion becomes incoherent. Moreover, because empirical, logical, and conceptual relationships can always either be coherently called into doubt or generate shallowness, we cannot reason in the third person omniscient voice. We cannot, as it were, climb out of our skin to talk about the world. Given all this, Harman’s argument for infinitely withdrawn objects seems to suffer from three interconnected problems that may prove fatal: 1. 2. 3. It fails to identify or establish an indispensable relation, thus leaving an insurmountable gap between antecedent practices and consequent ontological commitments. Instead, it relies on logical and conceptual relations, and is therefore shallow. The shallowness of the argument also involves an illicit universalization. These problems arise from Harman’s attempt to critically appropriate the ‘as-structure’ uncovered by Heidegger’s phenomenological analysis of tools. Now, within the compass of Heidegger’s thinking, the tool-analysis does in fact instantiate an indispensability argument. Just look at how circumspection is supposed to operate.14 Harman’s argument for withdrawn objects, by contrast, begins abstractly, and treats Heidegger’s analysis as merely vorhanden. 15 Those familiar with Harman’s Tool-Being will immediately recognize the crucial point. Consider the following line of reasoning: “The being of these objects and all others,” Harman writes, “consist[s] in the execution [Vollzug] of their reality prior to any coming-to-presence. The hammer must be Vollzug before it can be fertig; equally, the eyeball must be Vollzug before it can be fähig. Or rather, both Fertigkeit and Fähigkeit consist in the primordial execution of their being” (TB 72). Harman’s point here is entirely conceptual: enactment or execution is logically prior to readiness or capability. And he arrives at this conceptual point independently of any practice, first person experience, or context. By concentrating on the concept of enactment, Harman simply suggests that we extend it to objects in general, because there’s no immediate inconsistency in doing so. But this means that Harman dissolves the indispensability relationship binding practice, projection and context together, in favour a purely conceptual form of analysis. Without an indispensability relation, Harman’s argument cannot but become shallow. It rests entirely on a material implication: If Heidegger is right about equipment, then every object 7 Alexei Procyshyn The New School For Social Research withdraws. As we have seen, this conditional is vacuous, because it is shallowly grounded by a concept. The only connection one can establish between antecedent and consequent here is conceptual, logical, or analytic, and so Harman’s argument devolves into mere terminology. On the basis of mere concepts, Harman simply has no way to move from a hypothetical statement concerning equipment to a metaphysical commitment. The shallowness of Harman’s argument entails that his generalization of Heidegger’s asstructure is illicit, for it violates the restrictions governing universal introduction. As we know, introducing a universally quantified assertion must respect two rules for ensuring that the instantiating constant of an assertion actually names a whole class of entities, and not just a specific individual. This requires us, first, to ensure that the instantiating constant we now want to generalize does not occur in an open (non-discharged) assumption; second, the instantiating constant cannot also be present in the universalized statement. Harman’s attempt to universalize withdrawnness runs afoul of both restrictions. Given the conditional form his argument takes, ‘withdrawnness’ is an open assumption. For Harman simply assumes Heidegger’s account of withdrawnness-relative-to-Dasein, without discharging the assumption (i.e. proving it to be true). In fact, Harman’s decisive arguments all involve first-person instances of withdrawal in subtle ways. Harman’s generalization thus effectively misconstrues the universe of discourse in which the argument unfolds, and takes the properties of a specific kind of relationship to be a property of objects as such.16 To demonstrate the pervasiveness of the problem, consider Harman’s summary of his own argument in Tool-Being (220-225), from which I have excerpted for you the following passage: Let’s imagine […] a bulky metallic appliance is abandoned on a frozen lake. […] [S]ome sort of determinate encounter clearly does occur between them. This soulless piece of metal certainly does not enjoy immediate contact with the tool-being of the lake, as if sheer causal proximity were sufficient for capturing that lake in its withdrawn execution […]. Even in this case, the appliance reacts to some features of the lake rather than others – cutting its rich actuality down to size, reducing it to that relatively minimal scope of lake-reality that is significant to it. Note that the tool-being of the lake comprises an indefinitely large array of features, most of them irrelevant to the object lying on the surface. Simplifying somewhat, we can say that the stove reduces the lake to the single aspect of a frozen surface, to sheer “equipment for remaining stationary.” This sort of analysis is familiar enough to readers of Heidegger, at least in a human context. When the lake supports the appliance, this act of supporting unfolds entirely within the as-structure, not within the kingdom of tool-being. […] I hold that the resulting interaction between stove and ice is philosophically identical with the more familiar case of Dasein and the broken hammer. For what is decisive in the famous account of the “broken tool” is not that implicit reality comes into view, as if human surprise were the key to the reversal within being. Rather, the important factor is that the heavy object, while resting on the ice as a reliable support, did not exhaust the reality of that ice. (TB 222-223) With our keener sense of what’s involved in transcendental argumentation, many of the problems emerge of their own accord. As his references to Heidegger’s analysis illustrate, Harman’s thought experiment assumes a first-person perspective of the as-structure, and models the relationship between appliance and lake by analogy. Moreover, the discussion is limited to what we can perceive of the interaction, and does not consider ‘the interaction as it occurs in-itself’; the justification for the analogy is simply that we can – indeed, for the purposes of a coherent explanation, need to – perceive no more than this simplified interaction. Harman then attempts to show that the logic underwriting our first person perspective generalizes. But his 8 Alexei Procyshyn The New School For Social Research generalization simply repeats the particulars in the subjunctive mode of imagination. Does the counterfactual scenario actually transport us from the provinciality of our first-person perspective and explanatory biases to the metropolis of infinitely withdrawn objects? I am unconvinced. First of all, it remains totally unclear why the tool-beings of the lake and appliance are not exhausted by the complex causal interactions they are engaged in: other than Harman’s denial we have nothing to go by.17 His point is merely that our explanation of the event requires a foreshortened set of interactions since we are unable to perceive the entirety of the causal interaction anyway. But the sufficiency of causal explanation and the limits of human perception are inadequate grounds for ontological commitments. Indeed, explanation is relative to context, interest, and level of comprehension, and a subjective limitation is utterly insufficient to generate any ontological commitment. We can synthetically reformulate these problems into a full-blown contradiction: if we successfully identify an indispensable feature of a practice that ontologically commits us to the withdrawal of equipment, we cannot generalize withdrawal, since it is explicitly indexed to a human practice. If we disavow our first-person perspective, we cannot motivate the idea of withdrawal in the first place; worse, we are forced to smuggle some such perspective right back into our thought experiments, and this undermines our initial stance, since the perspective turns out to be indispensable after all. This is where Harman’s entire project founders, I think. Still, contradiction aside, he has nevertheless identified a fundamental problem for us. As he explains in Guerilla Metaphysics, “It is not […] possible to get ‘closer’ to […] things in such a way that their presence could provide some sort of measuring stick for how nearly we approached reality” (86). That is indeed the sceptical problem we need to address, and the source of the monistic and dualistic demands we are labouring under. As Harman puts it, practices like “my own loose understanding of how the sun functions, [or] a Stone Age shaman’s worship of it and an astrophysicist’s deep grasp of solar reality are quite distinct form one another” (86) and generate different ontological commitments. So far no disagreement; where we part ways, however, is on how to understand the relationship between entity, ontological commitment and human practice. Harman, as I’ve argued, opts for a conceptual analysis that leads him to over-extend the grammar of ‘object’ into a metaphysics of infinitely withdrawn, or vacuum sealed entities. Though undoubtedly subtle and sophisticated, my sense is that these reflections remain shallow: they lead him to an ontology without commitments, or a metaphysics without import, precisely because they do not distinguish between commitment and entity, meaning and reference. 9 Alexei Procyshyn The New School For Social Research 1 Harman, Graham, Guerilla Metaphysics: Phenomenology and the Carpentry of Things (Chicago: Open Court Press, 2005), p. 97. Cited hereinafter as GM. 2 So understood, my view of metaphysics is similar to the neo-Aristotelian view proposed by Jonathon Schaffer. See his, “On What Grounds What,” Metametaphysics, ed. David Chalmers, David Manley, and Ryan Wasserman (Oxford: Clarendon Press, 2009), pp. 347-383. 3 My discussion here draws on David Manley’s Introduction to Metametaphysics (pp. 1-37). 4 A historical example of shallow metaphysics would be what Karl Ameriks has called ‘short arguments to Idealism.’ See Karl Ameriks, Kant and the Fate of Autonomy: Problems in the Appropriation of the Critical Philosophy (Cambridge: Cambridge UP, 2000), esp. pp. 163-86. Short arguments originate as a meta-theoretical response to a skeptical worry concerning the legitimacy of a specific form of epistemological limitation. That is, they emerge from the need to ecumenically establish the truth of some claim independently of a set of disputed background commitments. To this extent, short arguments are not constructive; they do not set out to elaborate and defend a new theoretical position, but to justify or lend further credence to an already instituted thesis. As such, short arguments are derivative, stipulative, and finally trivial. Moreover, depending on the precise formulations involved in the stipulative definitions, short arguments may also be viciously circular. As Ameriks explains, “The key idea of what I mean by a ‘short argument’ is, very roughly, that reflection on the mere notion of representation, or on such very general features as passivity or activity involved in representation, is what is meant to show that knowledge is restricted from any determination of things in themselves” (163). According to Ameriks’ reconstruction, the short argument arises in response to a perceived failure in Kant’s work to show that the empirical realm a priori exhausts all that we can know theoretically. That is, the origins of the short argument are inextricably bound up with arguments concerning the nature of totality, which is delimited by the things in themselves, which Kant’s original readers believed was essential for demonstrating that the First Critique exhausted the limits of possible knowledge. The necessity attaching to the ‘thing in itself,’ they seem to have thought, guarantees the exhaustiveness of Kant’s theoretical framework. Like the transcendental arguments we have been considering, what is at issue is the source of this necessity. However, instead of a progressive argument to establish that a given feature of our subjective experience is indispensable, short arguments proceed analytically. What is problematic about them is that they are completely trivial in the sense described above – they simply define a skeptical worry out of existence (130), without sufficiently grounding the definitions they use. 5 A more technical explanation for this paradox has to do with existential import. According to Aristotelian logic, the subject position of the judgment form ‘S is P’ cannot be empty. This explains why the traditional square of opposition allows one to move from a universally quantified assertion (A: ‘All humans are mortal’; E: ‘No Gods are human’) to their respective subalterns (I: ‘Some humans are mortal’; O: ‘Some Gods are not human). Contemporary logic demonstrates why existential import is problematic: universally quantified assertions turn out to be material conditionals, and the antecedents of material conditionals can be false (i.e. have empty extension). Hence it is possible for an A statement to be true (albeit a ‘vacuous truth’) and its subaltern false. Hence, the assertion ‘No gods are human’ is vacuously true, whereas ‘Some gods are not human’ is false because there are no gods. Substance-based metaphysics tends to operate as if Aristotelian logic were still the appropriate means of analysis. 6 This peculiarity is well known. Russell and Whitehead make note of it in the Principia Mathematica (vol. 1: p 99, and C.I. Lewis gave it what I take to be perhaps the best chracterization. “The development of the algebra of logic,” he writes at the beginning of his 1912 paper published in Mind (vol 21, pp. 522-31), ‘Implication and the algebra of logic,’ “brings to light two somewhat startling theorems: (1) a false proposition implies any proposition, and (2) a true proposition is implied by any proposition” (522). 7 Paul Franks, All or Nothing: Systematicity, Skepticism, and Transcendental Arguments in German Idealism (Cambridge: Harvard UP, 2005), cited hereinafter as AN; Michael Williams, Unnatural Doubts: Epistemological Realism and the Basis of Scepticism (Oxford: Blackwell UP, 1991), cited hereinafter as UD. 8 Hence philosophy’s perpetual oscillation between realism and skepticism. Although it may sound as if the relationship between epistemology and metaphysics is inherently foundational, my own sense is that foundationalism is but one possibility. Various holisms (or coherentisms) are possible too. It is, I think, a matter of relative theoretical freedom whether we ultimately opt for the Heideggerian ontic-ontological circle, Quine’s naturalism and holism, or Kant’s Transcendental Idealism, etc. At bottom, what matters is how well a given ontology grounds our practices, not whether it is foundationalist. 10 Alexei Procyshyn The New School For Social Research 9 Robert Stern, Transcendental Arguments and Scepticism: Answering the Question of Justification (Oxford: Clarendon Press, 2000), p. 8. Stern actually identifies four kinds of transcendental arguments and a number of distinct strategies that respond to very specific skeptical worries. Stern’s taxonomy of transcendental arguments runs as follows: 1. 2. 3. 4. A is a truth-directed transcendental argument, where X is specified as some non-psychological fact or state of affairs which is claimed to be a necessary condition for experience, language, etc. e.g. For experience to be possible, there must be physical objects. A is a belief-directed transcendental argument, where X is specified as some belief which is claimed to be a necessary condition for experience, language, or some other belief(s), etc. e.g. For individuals to have beliefs about their own mental state, they must believe that there is an external world. A is an experience-directed transcendental argument, where X is specified as a way in which things must be experienced as being or appear to be, as a necessary condition for having experience of another kind, or language, beliefs, etc. e.g. For individuals to have subjective sensations or feelings, they must have experience as of an external world. A is a concept-directed transcendental argument, where X is specified as a context in which a concept-user must have acquired the capacity to employ the concept C, as a necessary condition for acquiring the capacity to apply the concept C at all. e.g. For individuals to have learnt how to apply the concept 'pain’, they must have acquired the capacity to apply that concept to others as well as themselves. (10-11) Truth-directed transcendental arguments, it turns out, always fail, because they conflate ontological commitment and ontological entity. Truth-directed, or “world-directed” transcendental arguments – as Quassim Cassam also calls them in “Self-Directed Transcendental Arguments,” in Robert Stern (ed.), Transcendental Arguments: Problems and Prospects (Oxford: Clarendon Press, 1999), pp. 83-110 – fail to coherently move from what one must believe to the way the world actually is. The other three types can be employed to combat the various forms of skeptical doubt. As Stern goes on to show, these four argument-types can be used to respond to either a justificatory skepticism (which further comprises reliabilist and normativist sub-types), or an epistemic skepticism. And, in combination with these skeptical stances, Stern articulates another ten strategies (for his full taxonomy, see 124). Although immensely important for understanding the contemporary senses of ‘transcendental argument’ and the contexts of their uses and assessments, Stern’s taxonomy is too refined to help us here. Without wishing to engage in a guerrilla attack against his efforts, I must admit that I side with Franks and Williams and maintain that transcendental arguments are a response to the Agrippan trilemma, and hence respond to justificatory skepticism. Moreover, insofar as justificatory skepticism aims to impugn our epistemic relation to the world, I do not see what motivates Stern’s distinction between these two skeptical types. The distinction strikes me as verbal rather than substantive. Hence I find his division of skepticism into three branches suspect as well as his categorization of our immanent critical strategies that follows from it. It might also be worth mentioning that Meillassoux’s main thesis in After Finitude: An Essay on the Necessity of Contingency (trans. Ray Brassier [London: Continuum Press, 2008]) hinges on critiquing this kind of two-step argumentative strategy, and the idea of necessity at work within it. Meillassoux’s argument, however, founders on his confused understanding of ‘necessity’ itself. For, in the first instance, he fails to distinguish between four kinds of necessity: on the one hand, he ignores the fundamental distinction between de re and de dicto necessity (most evident in his rejection of the principle of sufficient reason); on the other, he fails to address Kripke’s fundamental distinction between a posteriori necessity and its a priori counterpart. Worse, he even tends to conflate contingency with arbitrariness. Meillassoux thus utterly fails to see that an a posteriori necessity remains necessary for our actual world – i.e. it is simply not strongly rigid (i.e. does not pick out an existing entity in every possible world) – even though things could have been utterly different (contingent in his sense), while fundamentally confusing the de dicto necessity of explanatory theories (e.g. the formulation of projectible laws) with an ontologically realist commitment to the necessity of those laws existing in nature (de re). To address this naked assertion properly would take us well beyond the scope of our present discussion, since it would require us (1) to explicate how Meillassoux’s argument from ancestrality seeks to render the basic distinction between monistic and 11 Alexei Procyshyn The New School For Social Research dualistic demands inoperative (After Finitude, 1-27), and (2) how such an effort fails. Nonetheless, let me point out where one would find the means necessary to do so: Meillassoux’s principles of unreason and factiality. ”We are no longer upholding a variant of the principle of sufficient reason,” Meillassoux explains, “according to which there is a necessary reason why everything is the way it is rather than otherwise, but rather the absolute truth of a principle of unreason. There is no reason for anything to be or to remain the way it is; everything must, without reason, be able not to be and/or be able to be other than it is” (60). If theory and metatheory are still taken to be operative, Meillassoux’s claim is simply incoherent, because it fails to distinguish between an explanatory standpoint and the theory this standpoint valuates. It collapses causal and analytic necessity into metaphysical necessity in order to show that causal and analytic (mathematical) interactions presuppose a given configuration of entities and events, which cannot be explained by appeal to the same kind of necessity – hence appearing to be arbitrary, since no contingent causal history can elucidate the metaphysical necessity behind things being thus and so. By failing to distinguish between first- and second-order discourses, Meillassoux shoehorns ‘contingency’ into the same category as ‘arbitrariness.’ Despite implying that things could have been otherwise, the meaning of ‘contingency’ does not entail anything as metaphysically volatile as Meillassoux’s notion of unreason. More sympathetically phrased, although Meillassoux is right to distance himself from Leibniz’s and Spinoza’s speculative identity (read ‘conflation’) of Reason, Causality, and Necessity, he simply fails to see that the negation of sufficient reason does not lead to chaos (see After Finitude 71 for this mistake). Consider a few standard examples: ‘All bachelors are unmarried men.’ There is no doubt a necessary relationship between the meaning of ‘unmarried man’ and the meaning of ‘bachelor.’ Now this necessary, analytic relationship is contingent on how the meanings of these two terms were fixed. But the contingency of linguistic drift is neither chaotic, nor absolutely arbitrary. Put slightly differently, although analytic statements are necessary within a given conceptual framework, and ‘a priori,’ these frameworks and their semantic development of the terms is not. There can be no such thing as an a priori genealogy or etymology of a concept. To be sure, the Leibnizian and Spinozist think such an approach is perfectly valid. But showing them to be wrong in this respect does not in any way commit us to the hypertrophied thesis Meillassoux champions. Similarly, should it turn out to be true that all animals with lungs also have kidneys, this necessary relationship would still be contingent upon the historical facts of evolution on Earth. But the contingency of these facts is not tantamount to the arbitrariness of an organism’s genesis. Although it may have been possible for a being to come into existence that has kidneys and no lungs, such a being cannot appear ex nihilo, as a product, perhaps, of an unconstrained and irrational demiurge’s whim. Although any number of chance occurrences could have led to a completely different organism, the latter’s causal genesis is not arbitrary, even though it is contingent. More simply put, Meillassoux appears to believe that, unless something is necessarily designated across all possible worlds (i.e. some entity is strongly rigid), it is inherently arbitrary. Clearly that cannot be the case. All of these issues, moreover, manifest themselves in Meillassoux’s equivocal sense of reason (sometimes meaning ‘explanation’ or ‘epistemological justification,’ sometimes meaning ‘cause’ or ‘metaphysical ground’). And his principle of unreason simply misunderstands what reasons are (explanatory reconstructions of some first-order domain of research). 10 The example belongs to Barry Stroud. See Barry Stroud, “Transcendental Arguments,” in The Journal of Philosophy 65.9 (1968): 241-56. Reprinted in Barry Stroud, Understanding Human Knowledge (Oxford: Oxford UP, 2000), pp. 9-25; cited hereinafter as UHK. All references to Stroud’s work will be to the latter book. 11 See also, W.V.O. Quine, ‘”Things and Their Place in Theories,” Theories and Things (Cambridge: Belknap Press, 1981), pp. 20-22. 12 As Mark Colyvan (Indispensability Arguments in Mathematics [Oxford: Oxford UP, 2001]) has shown, Quine’s indispensability arguments depend upon two commitments: his peculiar brand of naturalism and confirmational holism. By ‘Naturalism’ Quine simply meant that there is no first philosophy and that our ontological commitments should follow form our best (scientific) practices rather than be dictated by aprioristic concerns. According to Colyvan, Quine’s “Naturalism tells us (1) we ought to grant real status only to those entities of our best scientific theories and (2) we ought to (provisionally) grant real status to all the entities of our best scientific theories. […] I’ll call this first strand of Quinean naturalism the no-first-philosophy thesis” (23). By confirmational holism, Quine understands something like Franks’ ‘Absolute grounding.’ The idea is that a given claim confirms or disconfirms the whole body of a theory rather than some isolated feature of it. A claim is therefore grounded in a system. Confirmational holism stresses that “it’s the whole body of theory that is tested, not isolated hypotheses” (35). 13 AN 246-48. 14 See Being and Time 68f. Handiness, as Heidegger insists, is always relative to Da-sein’s projection. “Strictly speaking,” he writes, “there ‘is’ no such thing as a useful thing. There always belongs to the being of a useful thing 12 Alexei Procyshyn The New School For Social Research a totality of useful things in which this useful thing can be what it is. A useful thing is essentially ‘something in order to …’ […]. The structure of the ‘in order to’ contains reference of something to something […] A totality of useful things is always already discovered before the individual useful thing” (68). Like the absolute grounding within the system of grounds that Franks makes central to refuting the Agrippan trilemma, Heidegger’s totality grounds the usefulness of an individual entity by indexing the tool’s indispensability to the practices we happen to be engaged in. “The act of hammering,” as Heidegger puts it, “discovers the ‘handiness’ of the hammer” (69). And all of this dovetails in Heidegger’s notion of circumspection: “Our association with useful things is subordinate to the manifold of references of the ‘in-order-to.’ The kind of seeing of this accommodation to things is called circumspection” (ibid.). 15 Indeed, Harman even takes Heidegger’s text to be ‘broken’ in virtue of its narrow, anthropocentric focus… 16 For a detailed account of the problem, see Merrie Bergmann et al, The Logic Book, 5th ed. (Boston: McGraw-Hill, 2009), pp. 532-542. 17 Harman displays a distressing tendency to simply deny contrary interpretations, without offering any explanation – let alone justification – for his dismissal. At the beginning of his analysis of Heidegger’s ‘tool,’ for instance, he writes, “A tool exists in the manner of enacting itself; only derivatively can it be discussed or otherwise mulled over. Try as hard as we might to capture the hidden execution of equipment, we always lags behind. There is no gaze capable of seizing it, despite Heidegger’s claims to the contrary” (TB 22). Other than Harman’s word, I have no clue why Heidegger’s claims on the matter shouldn’t be considered seriously. Similarly, we find him telling us, on the first page of Guerrilla Metaphysics, that “[t]his book […] rejects both analytic and continental traditions. […] In fact, another term that might be used for object-oriented philosophy is guerrilla metaphysics – a name meant to signify that the numerous present-day objections to metaphysics are not unknown to me, but also that I do not find them especially compelling” (1). We can partly attribute these rejections to the verve of a talented writer. Nonetheless, I would very much like to know which criticisms Harman is aware of, and have him explain why he is not bothered by them. Not everything in the history of philosophy is interesting, but those elements that are immediately relevant to one’s project – critiques included – ought to be addressed. 13