Brassier - Nominalism, Naturalism, and Materialism - Sellars's Critical Ontology (Chapter 7 from Contemporary Philosophical Naturalism and Its Implications)

7 Nominalism, Naturalism, and Materialism Sellars’s Critical Ontology Ray Brassier Nominalism denies the existence of abstract entities (universals, forms, species, propositions, etc.). Traditional nominalism proceeded from an empiricist epistemology that challenges the very possibility of metaphysics, whether idealist or materialist. The critique of empiricism is taken to entail the refutation of nominalism. But nominalism contains a valuable insight for naturalists: reality does not have propositional form. This is an insight that deserves to be taken up by post-Darwinian naturalists, for whom realism about abstract entities is problematic insofar as it seems to reiterate the theological presumption of a preestablished harmony between conceptual order and real order. The question is whether contemporary naturalists can take up this nominalistic insight while jettisoning the empiricist prejudices that tied it to skeptical relativism. For the claim that reality is devoid of propositional form need not require denying that we can capture aspects of reality in propositional form or that concepts have ontological purchase. The challenge is to explain both how propositionally structured thought arises within nature and how it can be used to track natural processes despite the lack of congruence between propositional form and natural order. Answering this challenge is central to Wilfrid Sellars’s understanding of what it means to be a post-Darwinian naturalist. Proceeding from a trenchant critique of the idea that the mind is a mirror of nature, Sellars’s stance exemplifies the ideal of what Huw Price calls “naturalism without mirrors.”1 But unlike his Rortyan descendants, Sellars refuses to relinquish the Enlightenment conviction that post-Galilean natural science marks a decisive advance in our species’ cognitive evolution. He defends a rationalistic naturalism that seeks to further the unfinished project of Enlightenment. Properly understood, his attack on the myth of the given disqualifies conceptual idealism and provides the precondition for an unapologetically transcendental realism. The key to the link between Sellars’s rejection of the given and his endorsement of realism lies in his metalinguistic nominalism. It is in the context of the latter that he insists that “propositional form belongs only in the linguistic and conceptual orders” (NAO, 62).2 My aim in this chapter is to explain why this insistence follows from Sellars’s rejection of the given and how it figures in his account of representation. Since my 6244-181-3pass-S3-007-r02.indd 101 8/22/2013 6:00:36 PM

102 Ray Brassier goal is primarily expository, I will not address the various objections that might be made against Sellars’s principal claims; others have done so already.3 I will begin by framing Sellars’s naturalist agenda within the context of what I call the post-Kantian problem of critical ontology. Then I will recapitulate how Sellars’s critique of the given determines his commitment to nominalism. This will be followed by a brief summary of Sellars’s functional role account of meaning. Lastly, I will discuss Sellars’s account of the relation between language and reality in terms of what I will call “methodological materialism” and indicate its link to Sellars’s adumbration of a metaphysics of processes. 1. THE PROBLEM OF CRITICAL ONTOLOGY Ontology is an attempt to answer the question “What is there?” But this cannot be answered by listing names of entities, for example, “table,” “chair,” “tree,” “Cyprus,” “Dante,” and “Aeroflot.” ‘Table,” “chair,” and “tree” are common nouns—that is, names for types of objects. “Cyprus,” “Dante,” and “Aeroflot” are proper nouns—that is, names for particular objects. Listing nouns, whether names of types or names of particulars, is uninformative because it offers us names without explaining what a name is or how it is related to its nominatum. If ontology is to take the measure of Kant’s critique of dogmatic metaphysics, it cannot remain content with conjuring yet another more or less arbitrary account of what there is; it must explain how we know what there is. One of the implications of epistemic finitude is that we do not know the divine names of things. Human names for things are not necessarily linked to the things they name. The meaning of a name is not the designation of an essence. Sellars accepts these Kantian strictures. Nevertheless, he will maintain that there is a sense in which nomination is at the basis of linguistic functioning. However, he insists that this nominative dimension does not pertain to meaning but to material process. The names in which linguistic signification finds its ultimate “aboutness” do not signify; they are material patterns. Thus Sellars’s solution to the problem of reference is at once dialectical and materialist: it ensures the autonomy of signification by grounding it in nonsignifying patterns. There are three basic desiderata for a critical ontology. It ought to explain • what a name is and how it relates to what it names; • why there is a difference between names and things; and • what kinds of things there are and indeed what kinds are. Thus the answer to the question “What is there?” also implies a response to the question “What is a category?” In its simplest version, the controversy over the status of categories is whether they are mind-independent attributes of one or several substances or mind-dependent concepts (Kant’s 6244-181-3pass-S3-007-r02.indd 102 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 103 pure concepts of the understanding or Heidegger’s existentials). Sellars’s suggestion is that they are neither. Categories are metalinguistic functions, but their metalinguistic function is nevertheless a mode of representing reality. The determination of categorial status depends on the identification of conceptual place. To specify an entity’s category will be to determine certain logico-semantic features of representation such that it is of that entity. Nevertheless, representation itself is not a conceptual relation or a relation between concepts and things. It is this account of the nonconceptual nature of representation that distinguishes Sellars’s from traditional accounts of representation as a relation between thoughts and things and renders his stance materialistic in a methodological rather than metaphysical sense. (I will try to elucidate this distinction below.) Because categories do not represent or designate any feature of the world, they are not phenomenologically intuitable. They cannot be read off the structure of language or of reality. To assume otherwise is to fall prey to the myth of the given. 2. THE MYTH OF THE GIVEN This myth has two facets: epistemic and categorial. The myth of the epistemic given is rooted in the confusion of thinking with sensing. It is crystallized in the following “inconsistent triad” of premises generated by empiricism: A. X senses red sense content S entails X knows noninferentially that s is red. B. The ability to sense sense contents is unacquired. C. The ability to know facts of the form x is ø is acquired. A and B together entail not-C; B and C together entail not-A; A and C together entail not-B.4 Sellars’s argument is complex, but it can be briefly summarized as follows. Knowledge is of facts (that such and such is the case). Facts have propositional form (x is ø). The question is whether we possess the capacity to sense facts. Either the ability to sense facts of the form x is ø is acquired or it is unacquired. If it is acquired, then it is not a sensory capacity, since by hypothesis, the ability to sense sense contents is unacquired. So the ability to sense facts must be unacquired. But if facts can be sensed, then sensation must have propositional form. To say that we sense facts is to say that sensation mirrors a reality already endowed with propositional form. But propositional form is tantamount to intelligible order. How then are we to explain the congruence between sensible order and intelligible order? If the ability to sense facts is unacquired, it cannot be explained naturalistically in terms of evolution by natural selection. Thus the congruence between 6244-181-3pass-S3-007-r02.indd 103 8/22/2013 6:00:37 PM

104 Ray Brassier sensible order and intelligible order must either be left unexplained or explained by invoking supernatural factors. Sellars argues that premise A is false. Sensing a red sense content S does not entail knowing noninferentially that S is red.5 Sensory awareness alone does not constitute knowledge. The noninferential knowledge that x is ø—seeing that blood is red or hearing that the clock has struck twelve—is a conceptually mediated perception, not a sensible intuition. The perceptual immediacy of such knowledge is mediated by an elaborate conceptual framework of objects related in a publicly observable space and time. This conflation of thinking with sensing, or in Kantian parlance, concepts with intuitions, leads straight to the second facet of the myth, the myth of the categorial given, according to which the categorial structure of reality (assuming it has one) impresses itself upon the mind the way a seal impresses itself upon wax.6 This is to conflate sensing with sensing as.7 But sensory awareness is not awareness as. To be aware of an item having categorial status F is not to be aware of it as F. To sense something as F is to deploy the concept of F. This deployment is rule governed. But rule following is thinking, which is an activity irreducible to sensing, even if bound up with it in the case of sapient beings. Thinking cannot touch the real: it belongs to a different order. But as we shall see, this does not oblige us to deny that we can successfully think about the real, or deny that thinking is necessarily embedded within reality. Sensation is of the real but cannot be about it. Thinking is about the real but cannot enter into direct contact with it. But before we say something about how we ought to understand thinking once we have abandoned the myth of the given, we must explain why rejecting the myth does not entail skepticism. It is important to see that skepticism buys into the myth of the epistemic given: it has to assume that appearances are given with determinate conceptual characteristics even as it questions whether there is any correspondence between the structure of appearance and the structure of reality. So skepticism unwittingly presupposes knowledge of appearances even as it presumes to cast doubt upon their connection to reality. Moreover, it cannot explain why there are appearances (for it cannot claim that appearances are mere appearances on pain of a problematic regress). But to assume that determinately structured appearances are given is already to assume too much. Once the myth of the epistemic given has been dispelled, it becomes necessary to admit that sensing as already presupposes knowledge of a structured domain of interrelated objects existing in a framework of space and time. Both thinking and sensing are oriented toward the world, albeit in distinct yet interrelated dimensions: in Sellars’s words, the of-ness of thought is not the of-ness of sensation.8 It may sound platitudinous to reiterate once more the claim that the mind is not a private inner sanctum, that it is externalized in the world, and that this externalization is a consequence of its connection to linguistic activity. Sellars shares Heidegger’s and Wittgenstein’s rejection of the Cartesian privacy of the mental. But unlike both of them, he situates this insight within a 6244-181-3pass-S3-007-r02.indd 104 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 105 broader metaphysical account of the place of mind in nature. That thinking and language use are both essentially activities will turn out to be crucial to this account. The concept of inner-thought episodes is modeled on publicly observable “sayings-out-loud.” This is a direct consequence of rejecting the myth of the epistemic given. The ability to apperceive our own mental states is acquired, not innate. Introspection is a corollary of extrospection. The ability to introspect and perceive that one is thinking X or feeling Y presupposes conceptual capacities rooted in linguistic practice. Yet this is not to say that thinking is reducible to verbal dispositions. The claim that speaking is prior to thinking in the order of knowing is compatible with the claim that thinking is prior to speaking in the order of being. Pre- or nonverbal thought is perfectly real. But our ability to apperceive ourselves thinking is linguistically mediated. Sellars is as much of a realist about inner-thought episodes as Descartes. His amendment to Cartesianism insists only that access to “inner” reality is just as mediated as access to “outer” reality. Sellars does not say that thoughts are necessarily public or even essentially publicizable; his claim is that our ability to understand what a thought is is tributary to communally generated and publicly shared conceptual resources. By the same token, Sellars’s realism about internal thought episodes does not commit him to the Cartesian claim that covert thought stands to overt speech as cause to effect. The claim that thinking is essentially related to the capacity for linguistic expression does not identify thinking with linguistic behavior. Rather, it postulates that the former is constitutive of the latter in the same sense in which molecules are constitutive of gases. Thoughts are “in” language-using animals in the same sense in which molecules are “in” gases.9 But it is just as mistaken to construe thoughts as the cause of linguistic behavior as to construe molecules as the cause of gas volume. The connection is one of constitution, not causation. Thus it is part of the definition of the occurrentness of thoughtacts that they be expressible in sayings-out-loud even if no saying occurs. Ultimately, empiricism and Cartesianism are not the only tributaries of the myth of the given. The claim that meaning is rooted in the originary “sense-bestowing” acts of consciousness renders phenomenology, at least in its transcendental variants, directly subservient to the myth. The moral to be drawn from the critique of the given is not just that our understanding of thought is modeled on our understanding of language; it is that the intentionality of thought derives from the intentionality of public discourse. Thought is not the locus of an originary intentionality that is subsequently transmitted to language. Intentionality is primarily a property of candid public speech established via the development of metalinguistic resources that allows a community of speakers to talk about talk. This rejection of the postulate of originary intentionality—a postulate embraced by philosophers as different as Husserl and Fodor—is another direct consequence of abandoning the myth of the given. But if intentionality is primarily a linguistic phenomenon, what does this imply for our understanding of meaning not only as it pertains to mind but also as it pertains to nature? 6244-181-3pass-S3-007-r02.indd 105 8/22/2013 6:00:37 PM

106 Ray Brassier 3. “MEANS” TALK Sellars defends a metalinguistic version of nominalism. Semantic statements such as “ ‘Rouge’ (in French) means red” quote a sign-design and correlate its function in an unfamiliar language with its function in a familiar language. More generally, semantic operators such as “means,” “designates,” “refers,” or “stands for” correlate the functioning of unfamiliar words with the functioning of familiar words. The “means” in a semantic statement such as “ ‘Rouge’ means red” is a special version of the copula. It establishes a correlation between a distributive singular term and a metalinguistic sortal (these technical expressions will be clarified below). Sellars uses the dot-quoting device to mark the exhibition of linguistic function: thus, “ ‘Rouge’ in F is a •red• in E” says that the mentioned sign-design plays the same linguistic role in French as “red” does in English. “Red” here is not mentioned but used in a special way, not as it is ordinarily used in English (as meaning the color red) but as an illustrating sortal in a metalinguistic assertion. It is this peculiar status that the dot-quoting device exhibits. In statements such as “A ‘rouge’ is a •red•” and “A ‘triangulaire’ is a •triangular•,” “rouge” and “triangulaire” both function as distributive singular terms rather than abstract nouns. This is to say that they function just as the expression “the lion” functions in the sentence “The lion is tawny.” In the latter, the property of being tawny is not predicated of an abstract entity, lionhood; rather, it is predicated of each particular10 lion: the singular term “the lion” refers distributively to particular lions existing in space and time. Similarly, at the metalinguistic level, one can treat the sign-designs “rouge” and “triangulaire” as distributive singular terms and establish a correlation with the metalinguistic sortals •red• and •triangular•. Doing so yields the correlation “The ‘rouge’ is the •red•” and “The ‘triangulaire’ is the •triangular•.” The next step is to write “ ‘Rouge’s are •red•s” and “ ‘Triangulaire’s are •triangular•s” in order to indicate that what is being correlated here are particular linguistic tokenings rather than abstract linguistic types. The pluralization shows that both distributive singular terms and metalinguistic sortals indicate patterns of tokenings rather than the names of abstract types. If this were not the case, we would simply be substituting reference to abstract linguistic entities for reference to abstract extralinguistic entities. In order for the nominalistic strategy to work, it is necessary to insist that only particular linguistic tokenings exist, not linguistic types. This is what is shown by correlating the distributive singular terms “rouge’s” and “triangulaire’s” with the metalinguistic sortals •red•s and •triangular•s. This metalinguistic strategy can also be used to explain why abstract singular terms should not be taken as designating abstract entities. Thus consider the abstract singular term “redness.” Sellars insists it is a mistake to treat it as the name of an abstract entity. The meaning of statements involving “redness” can be reformulated without loss using the predicate “red”: “A exemplifies redness” is equivalent to “A is red.” To claim that it 6244-181-3pass-S3-007-r02.indd 106 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 107 isn’t is to claim that the context “ . . . exemplifies redness” means something other than “ . . . is red.” But to say this is to insist that “exemplification” is also the name of an abstract entity: the “exemplification nexus” through which the object A is supposed to be connected to redness. It is important to see that to append this metaphysical annex to ordinary usage is to invite a regress, for if we insist on explaining the meaning of relational expressions in terms of designation, we will always require another abstract relational term to explain the relation between the particular, the universal, and their relation (this is a version of the Third Man argument). 4. PICTURING AS REPRESENTATION Yet Sellars’s nominalism is part and parcel of his realism. Indeed, he endorses a transcendental realism, albeit one that will be ultimately cashed out in terms of processes rather than objects. Thus what needs to be explained is the connection between metalinguistic sortals and extralinguistic reality. The crux of Sellars’s explanation of the link is his theory of picturing. The metalinguistic properties of sign-design tokens picture the nonlinguistic properties of objects. However, picturing is not a semantic relation. It is not to be understood as a correlation between elements in the conceptual order (the order of signification) and objects in the causal order (which for Sellars is nonsignifying), but rather as a nonconceptual correspondence existing within the natural-physical order. This correlation is what Sellars calls representation. Representation is not a relation between concepts and objects, but between objects and other objects. These objects are spatiotemporal particulars. Metalinguistic functioning is realized in the material properties of sign-design tokens. Thus the metalinguistic functioning of a sortal like •red• is correlated with a “real,” nonlinguistic physical property; but this property is not named by the sign-design “red”; instead, it is pictured by its syntactical role. Thus, in an expression such as “red a,” where the property “red” is predicated of the object named a, the token “red” does not stand for or designate the property. Rather, it is its syntactical concatenation with the name “a” that ultimately explains its connection to the extralinguistic property: In general the nominalist holds that there is no thing which “red” stands for because it stands for red things and red things is not a thing. Now it is a truism that the concatenation of “red” with “a” tells us that a is red. But is this fact illuminated by the idea that “a” is correlated with a and “red” with red things? I think not. To understand what is going on, we need a different perspective on the syntactical form of “red a.” I submit that it is properly viewed, not as a concatenation of two referring expressions, but rather as a name, “a,” which has the character of being concatenated to the left with a case of the sign-design (flatus vocis) “red.” If we abbreviate the expression 6244-181-3pass-S3-007-r02.indd 107 8/22/2013 6:00:37 PM

108 Ray Brassier Concatenated to the left with a “red” by red* we can say with respect to a token, t, of “red a” that while superficially t is a “red”^“a” i.e. a “red” concatenated with an “a”; its true form, its depth grammar, so to speak, is given by t is a red* [“a”] (Sellars MEV, 333) Thus Sellars’s suggestion is that metalinguistic properties picture real properties via the syntactical configuration of sign-design tokens. More generally, Sellars insists that the part played by relational expressions and empirical predicates in linguistic statements can be reconstructed without hypostatizing them as abstract entities.11 His chief inspiration here is Wittgenstein’s claim in the Tractatus that we say that aRb by placing the names “a” and “b” in a certain dyadic relation.12 This dyadic relation is a pattern of inscription. It is the inscription that shows how a and b are related by inserting the symbol “R” between the names “a” and “b.” But the relation itself is not an object. And the token “R” that relates “a” and “b” is not a name. Thus what “R” does in the statement “aRb” could be done without using a symbol. Consider the statement “a is larger than b.” We could adopt a convention whereby the graphic properties of the inscriptions “a” and “b” say what the statement “a is larger than b” says. For example: a b This inscription states what “a is larger than b” states without using the expression “is larger than.” But it is crucial to note that nothing in the above inscription plays the role (allegedly) played by “is larger than.” That “b” is below “a” is essential to the meaning of this statement. But this graphic feature does not correspond to the role played by the expression “is larger than.” Rather, in the inscription above, “b” ’s being below “a” plays the role played by “a” and “b” having “is larger than” between them. Thus both the “is larger than” and “be” ’s being below “a” are functioning here as inscriptions, which is to say graphic objects, rather than as signifying expressions. Similarly, the statement “x is red,” which means that object x has the property red, could be written x. Here it is the way in which the name “x” is inscribed that tells us what property the object x has. The inscription “x” has two relevant features: it features a token of the name “x” that refers to object x and it is has a specific graphic characteristic. Fundamentally, Sellars’s claim is that predicates do not play an independent role within linguistic expressions: “Not only are predicative expressions dispensable, the very function 6244-181-3pass-S3-007-r02.indd 108 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 109 played by predicates is dispensable” (NAO, 51). Thus it is a mistake to abstract the role played by predicates from the role of the expressions in which they occur. It is this abstraction of a fragment of function that encourages the mistaken idea that predicates designate conceptual properties or metaphysical attributes. The predicative role should not be reified and turned into an abstract entity called a “property” that exists independently of sentential contexts. Still less should the conceptual property supposedly expressed by the predicate be hypostatized and turned into an ontological attribute that exists not only independently of language—as conceptual properties are alleged to—but also independently of thought. As Sellars puts it, “The extralinguistic domain consists of objects, not facts. To put it bluntly, propositional form belongs only in the linguistic and conceptual orders” (NAO, 62). In the final analysis, conceptual functions are linguistically incarnated in sign-designs whose material characteristics picture objects. Semantic functions are nothing independently of this incarnation. But crucially picturing itself is not a semantic relation or function. Sellars describes it as a “secondorder isomorphism” between objects in the natural order: thus a CD pictures a piece of music via a complex transcoding of information from one physical medium into another. What lies at the core of Sellars’s nominalism is the idea that conceptual signification bottoms out in picturing. 5. NAMING AND PICTURING Sellars’s rejection of the myth of the given does not lead to linguistic idealism and the claim that reality is a linguistic construct. Rather, it is the cornerstone of a critical ontology in which language is embedded in a nonlinguistic, a-signifying reality that is devoid of propositional form. Thus Sellars’s nominalism is part and parcel of his embrace of naturalism and materialism (these are, of course, not equivalent). He is a naturalist because he claims that linguistic practices, in which thinking is rooted, are varieties of natural processes. Natural science investigates these processes. That linguistic activity is a distinct and possibly even a unique variety of natural process, whose specificity should not be elided, is part of Sellars’s Kantianism. What makes Sellars a materialist as well as a naturalist is his insistence that the varieties of natural processes extend well beyond those comprehended within the organic realm. To be a materialist is to refuse to organicize nature—that is, to use the organism as an explanatory paradigm for the whole of reality. What makes Sellars’s materialism nonmetaphysical is his insistence that while linguistic function is ultimately rooted in inorganic as well as organic patterns, these material patterns are perceptible, which is to say, “matter of factual” (in a sense to be clarified below): sounds, marks, spacings, movements. They constitute what Sellars calls “natural-linguistic objects.” Natural-linguistic objects are names that depict parts of reality. Names are part of the natural order but only insofar as they are meaningless. 6244-181-3pass-S3-007-r02.indd 109 8/22/2013 6:00:37 PM

110 Ray Brassier Just as reference is not a semantic relation, representation is not an epistemic relation: it is a natural function. A general theory of representational systems will distinguish between sensate and insensate varieties of representational function. This requires distinguishing propositional form from conceptual form. Animal representational systems operate through a propositional form that has both a referring and a characterizing aspect.13 Crucially, this propositional form is nonconceptual. Thus the representation of an object a’s being red is carried out by the characterizing inscription of a symbol token “a.” Both the referring and characterizing aspects of elementary propositions are syntactically encapsulated. Syntactic form is realized in the neurobiological properties of the nervous systems of sentient organisms. In this regard, propositional form is prelinguistic and more fundamental than logical or conceptual form. Logical form is inaugurated once representational systems can represent relations of association, compatibility, and incompatibility between representations. Metarepresentation establishes inferential relations between propositions. Full-blown conceptual form is achieved only once metarepresentation relays propositional structures endowed with the rich predicative resources of a natural language, which supervenes on but is irreducible to the coding procedures of a representational system. 6. PATTERNS AND PROCESSES At its most elementary level, Sellars suggests, linguistic practice is anchored in names that picture objects as being somehow. This somehow-ness is shown, not said, by the manner in which names are uttered or inscribed. An utterance or inscription by itself is not, of course, a statement. It is a physical pattern (phonemic, graphic, or gestural). Thus Sellars’s suggestion is that conceptual properties do not designate attributes or ways of being but that they are nevertheless rooted in acts of representing that picture reality in ways that can be said (from within the conceptual order) to be more or less adequate. The criterion of pictorial adequacy is formulated using our extant conceptual categories and, as such, is internal to our signifying scheme and dependent upon our available predicative resources, yet it can still be used to track the correlation between conceptual order and real patterns.14 Conceptual categories are embedded in and conditioned by natural function, even though they do not mirror the latter. Demonstrating this is the burden of Sellars’s philosophy. The demonstration requires distinguishing the propriety of conceptual function from any metaphysical correspondence between thoughts and things. Meaning is not a relation: meaning statements establish metalinguistic correlations between words and other words rather than a metaphysical relation between words and things. The basement level of language consists of pattern-governed connections between naturallinguistic objects and other physical objects. Words do not depict reality 6244-181-3pass-S3-007-r02.indd 110 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 111 because of what they mean but because of physical connections between the semantic regularities obeyed by speakers and the physical patterns in which these semantic regularities are embodied: The “real relation” which underlies the fact that “man” refers to men must surely be a real relation between the word “man” and men, a relation to be formulated in terms of generalizations having subjunctive form, which specify uniformities in which expression-tokens (including sentences containing the word “man”) and extra-linguistic objects (including men) are involved. (Sellars NAO, 61) These uniformities are incarnated in phonetic, graphic, or haptic patterns, as well as behavioral ones. They are exhibited in the uniformities of performance that constitute pattern-governed linguistic behavior. But these patterns reflect espousals of principle that constitute linguistic competence.15 Ultimately, categories are to be explained in terms of metalinguistic role. Metalinguistic role is to be explained in terms of correct representing. Correct representing is to be explained in terms of picturing. 7. CONCLUSION To what extent then does Sellars’s system satisfy the demands of a critical ontology? It proposes an answer to each of the questions I mentioned at the outset. These questions were, what is a name? How do names relate to the things they name? What kinds of things exist, and what are kinds? In answer to the first question, a name is a sign-design (i.e., a naturallinguistic object) with empirical characteristics (phonemic or graphic) whose tokenings are correlated with patterns of objects in accordance with what Sellars calls “ought-to-be” rules. In answer to the second question, a name relates to what it names because names are equivocal entities operating in two distinct but intimately connected dimensions: the semantic dimension and the material dimension. Names signify by virtue of their rule-governed linguistic role. But names are also a-signifying objects that picture other objects in the world through their sensible characteristics. Finally, the third question must be answered in reverse. First, kinds are metalinguistic sortals, and these in turn correspond to distinct patterns of rule-governed tokenings. As to what kinds “really” exist, the ultimate catalogue of “real” kinds will be identified by the absolute picture of the world, which is the regulative ideal of empirical enquiry. This is, of course, one of the most controversial aspects of Sellars’s transcendental realism and one that is disavowed by so-called left Sellarsians.16 Nevertheless, Sellars’s claim that we possess a criterion of adequacy for picturing is not quite so fatally unelucidated as critics have alleged. We can gauge our current picture’s 6244-181-3pass-S3-007-r02.indd 111 8/22/2013 6:00:37 PM

112 Ray Brassier degree of approximation to this ideal picture by measuring, in Jay Rosenberg’s words, “the absolute numerical magnitude of the correction factors which must be introduced into applications of the strict counterparts of predecessor laws to arrive at the values determined by their successors” (2007, 69).17 Moreover, we can orient our current picture toward this ideal by projecting categories into a successor theory through the construction of analogical models with counterpart properties.18 The outstanding question is how exactly predicative roles are exhibited by the material properties of the natural-linguistic objects in which all linguistic functioning is ultimately anchored. “For the perceptive nominalist,” writes Sellars, “the varieties of mapping [i.e., picturing] are as multiple as simple matter-of-factual qualities and relations” (NAO, 60). The key to Sellars’s “methodological materialism” lies in this claim that the dimensions of picturing vary with the varieties of matter-of-factual qualities. For it is the latter that provide us with our systems of coordinates for the material domain. Thus the logic of Sellars’s account implies that whatever conceptual properties are used to describe “materiality” remain provisional and subject to further adjustment and even fundamental revision. This fallibilistic aspect of philosophical ontology, which is implicit in the theory of picturing, distinguishes Sellars’s critical stance from that of more straightforwardly metaphysical—or “dogmatic”—materialists. Empirical theory, in Sellars’s words, is “a self-correcting enterprise,” but where more positivistic naturalists appeal to empirical evidence as the sole arbiter of theoretical revision, Sellars’s rationalistic naturalism grants a decisive role to philosophy. Its task is not only to anatomize the categorial structures proper to the manifest and scientific images respectively but also to propose new categories in light of the obligation to explain the status of conceptual rationality within the natural order. Thus philosophy is not the mere underlaborer of empirical science; it retains an autonomous function as legislator of categorial revision. It is in accordance with this legislative task that Sellars postulates the category of “pure processes” to explain the link between conceptual categories and material patterns.19 The rule-obeying activities constitutive of conceptual categorization and the pattern-governed behaviors in which they are embodied are distinct but correlative dimensions of natural process. Conceptual transformations track material patterns without mirroring them. Pure processes are postulated at the metacategorial level in order to explain the covariation between patterns of representings and patterns of represented objects. Yet this postulation is perfectly in keeping with Sellars’s commitment to methodological naturalism. It serves as a model that will be necessarily transformed in the course of its deployment by future empirical science. In this regard, Sellarsian naturalism is critical rather than dogmatic precisely insofar as it retains a role for a priori philosophical theorizing. However, the ontological categories first catalogued and then postulated by 6244-181-3pass-S3-007-r02.indd 112 8/22/2013 6:00:37 PM

Nominalism, Naturalism, and Materialism 113 philosophers are constrained by their explanatory role relative to empirical investigation and hence necessarily subject to future empirical revision. NOTES 1. Price 2011. 2. Sellars’s texts will be referenced using the abbreviations established by Jeffrey Sicha in his complete bibliography of Sellars’s philosophical works. 3. See deVries (2005), O’Shea (2007), Rosenberg (2007), and Seibt (1990, 2000). My understanding of Sellars’s thought is greatly indebted to all four. 4. Sellars EPM, 20. 5. Though, of course, this is not to deny that perceptual knowledge can be noninferential. Sellars wants to defend the legitimacy of noninferential knowledge at the level of perception by showing how it presupposes a background of inferential knowledge, which itself is not perceptual in character or sensory in origin. 6. Cf. Sellars FMPP, 12. 7. It would be instructive to compare and contrast Sellars’s views here with Husserl’s account of “categorial intuition.” 8. EPM, 55–56. 9. Ibid., 104. 10. I say “particular” rather than “individual” because the latter category is ambiguous between abstract and concrete individuals: a Platonist might regard lionhood as an abstract individual. Thus it is the spatiotemporal particular that is to be contrasted with the nonspatiotemporal universal here. 11. Sellars is careful to emphasize that his dispensability claim applies only to empirical predicates; cf. NAO, 51. 12. Tractatus 3.1432. 13. See Sellars MEV, 336. 14. See Seibt 2000. 15. In Sellars’s words, “Espousals of principle is reflected in uniformities of performance” (TC, 216). This is what James O’Shea calls Sellars’s “norm/nature metaprinciple” (O’Shea 2007, 62, and passim). 16. The coinage is Richard Rorty’s and can roughly be taken to designate those philosophers who embrace Sellars’s critique of the given while rejecting his commitment to scientific realism. 17. See also Johanna Seibt: Even if we cannot attain a framework-external standpoint and cannot discern from our present vantage point the framework in the limit from which predecessor frameworks deviate at arbitrarily small amounts, we can confirm the convergence of the series of frameworks on the basis of the Cauchy criterion of convergence, namely, by showing that the members of the series get arbitrarily close to each other. In this way, Rosenberg claims, while we have no descriptive characterization of the limit framework, we can compare the correction factors of the frameworks in the series and thus affirm from a framework-internal vantage point that there is a limit framework (2000, 264). 18. See, in particular, Seibt (2000) for an extremely illuminating account of the importance of “projective metaphysics” for Sellars. 19. See Sellars FMPP. 6244-181-3pass-S3-007-r02.indd 113 8/22/2013 6:00:37 PM

114 Ray Brassier REFERENCES Note: I follow the now standard convention of referring to Sellars’s texts using the acronyms established by Jeffrey Sicha in his complete bibliography of Sellars’s writings, “The Philosophical Works of Wilfrid Sellars.” This can be found in Sicha’s edition of Sellars’s Cassirer lecture notes, Kant’s Transcendental Metaphysics (Atascadero, CA: Ridgeview, 2002), 485–92. Sicha’s bibliography incorporates corrections by Andrew Chrucky, whose site Problems from Wilfrid Sellars also features a complete bibliography: http://www.ditext.com/sellars/bib-s.html. Works by Wilfrid Sellars Sellars, Wilfrid. Empiricism and the Philosophy of Mind [EPM]. Cambridge, MA: Harvard University Press, 1997. ———. “Foundations for a Metaphysics of Pure Process” [FMPP]. The Monist 64 (1981): 3–90. ———. “Mental Events” [MEV]. Philosophical Studies 39 (1981): 325–45. ———. Naturalism and Ontology [NAO]. Atascadero, CA: Ridgeview, 1996. ———. “Truth and ‘Correspondence’ ” [TC]. In Science, Perception and Reality. Atascadero, CA: Ridgeview, 1991, 197–224. Works by Other Authors deVries, Willem. Wilfrid Sellars. Chesham: Acumen, 2005. O’Shea, James. Wilfrid Sellars: Naturalism with a Normative Turn. Cambridge: Polity Press, 2007. Price, Huw. Naturalism Without Mirrors. Oxford: Oxford University Press, 2011. Rosenberg, Jay. “The Elusiveness of Categories, the Archimedian Dilemma and the Nature of Man: A Study in Sellarsian Metaphysics.” In Wilfrid Sellars: Fusing the Images. Oxford: Oxford University Press, 2007, 47–77. Seibt, Johanna. Properties as Processes: A Synoptic Study of Wilfrid Sellars’ Nominalism. Atascadero, CA: Ridgeview, 1990. ———. “Pure Processes and Projective Metaphysics.” Philosophical Studies 101 (2000): 253–89. 6244-181-3pass-S3-007-r02.indd 114 8/22/2013 6:00:38 PM