Brandom 1/14/19Sellars Week 5 (“Naming and Saying”) Notes1. LW in Tractatus: Two alternatives, either complex objects represent (don’t say ‘name’) facts,or facts about names do. Facts vs. complex objects. Jumblese.2. “Of particular interest in this connection would be the Jumblese translation of Appearance and Reality.” [105] Q: Why? A: Because of his regress-of-relations argument (compare: Lewis Carroll): If we say that R relates a and b, then we can ask how a and b are related to R.Let us suppose the answer is R’, a matter of their standing to the relation as its relata. But what is the relation between relations and their relata? In particular, how are a and b related to relation R’, the relation of being relata of a relation? Call this relation R’’. This is a vicious regress, Bradley argues, because we can’t understand any of the earlier relational statements without already understanding the later ones. He takes it to show that the idea of external relations is incoherent. WS is claiming that this argument cannot be formulated in the perspicuous language of Jumblese. Indeed, at [122] WS says what he means by “perspicuous language” for the purposes of this essay: “in a perspicuous language, i.e. a language which had a built-in protection against Bradley’s puzzle…”. 3. Discuss world-of-facts (LW, me) vs. world-of-objects (WS). One cost of former is that facts are about objects. So one needs an objective, mind-independent notion of aboutness. I think one can get that by substitution. One cost of the former is that one must treat facts as arrangements or configurations of objects, and that only seems to work for (at most) atomic facts. All others will need to be given some other analysis. And monadic facts seem particularly troubling. The issues about universals—nominalism, realism, conceptualism—are related to this question in complex ways.4. Sellars’s idealism: Sellars seems to think that all universals, all properties, all repeatables (cf.psychological nominalism) are ultimately to be understood in terms of linguistic functional repeatables. This is also a pragmatism.5. WS starts off the essay by agreeing with and enforcing LW’s point that facts picture facts. But in the end, Sellars does not believe in facts, at least, not in the pictured ones, only in the picturing ones—the linguistic ones. That is the ‘idealism’ I accuse him of (take him to endorse) in (4) above.6. What I’m shaping up to at this point is the accusation that Sellars is implicitly depending on far too simple a picture of the relation between pragmatics and semantics—or a failure to appreciate that his arguments address the former, but he immediately draws conclusions about the latter, conclusions that follow only given some very contentious collateral premises. For he has a nice account of what we are doing when we use various expressions—an account he describes, in a way that systematically blurs the distinction between tmpej1grptd 1 1/14/2019Brandompragmatics and semantics, as “metalinguistic”. But he concludes immediately (or at least without further argument or discussion) that because we are doing that, we are not (also) describing how things are. We can agree that we are not describing things in the narrow sense, just because there is also a “metalinguistic” pragmatic component that is absent in the paradigmatic ground-level case of basic matter-of-factual empirical descriptions. But he seems to think that that means that there is not also a wider notion of description (indeed, empirical description) that these cases (modal, ontological) fall under—at least, not any that is not as broad as saying, understood in declarativist terms. I think there is middle ground between (narrow) descriptivism and declarativism. 7. Green/triangular: a) Sellars’s point that color examples are bad, because there is only one word, ‘green’, to do the two jobs needed in i) a is green, and ii) green is a color (property), which invites the thought that since in (ii) ‘green’ is apparently and arguably functioning as a name, it must be doing so in (i) as well, while there are distinct words in iii) a is triangular and iv) triangularity is a shape (property), which does not invite that confusion. Thinking that the adjective in (i) and (iii) is a name in turn invites the exemplification reading of Bergmann. b) re Bergmann: discuss the sense in which his view is heterodox, in that he avoids the Bradley regress, but by stopping one move up, insisting that exemplification, and only exemplification, is perspicuously thought of as shown but not said—the same view Sellars (agreeing, he claims, with LW) says holds of all predicates, one level down. (This is the Saul/Paul, but a heterodox Saul passage.)8. Facts: WS agrees with LW that it is the fact that names are so configured that pictures (represents, expresses) the fact that objects are configured thus. But both facts must be treated metalinguistically. The former, representing, fact, must be metametalinguistic, since it is already, on the face of it, metalinguistic, in a strict, mentioning expressions not using them way, not the different—at best falling under a broader notion of of metalinguistic—pragmatic one that applies to the latter, represented fact.9. Discuss the semantic point at the end:a) Analogy between “the fact that p” and “that p is true.”b) Don’t see a distinct semantic fact in the world about truth. Rather, that fact is metalinguistic on its face.c) In fact, WS reads T-sentence equivalences as displaying facts about the activity of inferring, as expressing a commitment to do something. This is an interesting view on itsown.d) It is also connected to the stuff about conceptual change (which I did not get around to discussing) from the end of CDCM.10. Explicitly discuss both a) the threat of a meta-meta-meta…linguistic regress, given WS’s i) metalinguistic treatment of modality, universals, ‘truth’-talk… and ii) his classification undertmpej1grptd 2 1/14/2019Brandomfunctional-normative universals (does it matter that grammatically they are DST’s? Presumably not: “the lion” is going to have to be treated metalinguistically itself) account of what we are doing metalinguistically when we talk about properties and facts, apparently in the object language. And b) The question of whether progress is made and the regress is