I 199
Conservativeness/Expansion/Language/Tonk/Brandom: pro conservative expansion: if the rules are not inferentially conservative, they allow new material inferences and thus change the contents that were associated with the old vocabulary expressive logic/Brandom: requires that no new inferences that only contain old vocabulary be rendered appropriate by this (if they were not before). >
Conservatity.
I 200
E.g. "boche"/Dummett: non-conservative extension, statements that do not (!) contain the expression might now be inferred from others that do not contain it either E.g. inference from German nationality to cruelty BrandomVsDummett: this is not about non-conservatism: it only shows that the expression "boche" has a content which is not contained in the other expressions E.g. the cocnept "temperature" has also changed with the methods of measurement. It's not about novelty of a concept, but undesirable inferences. >
Concepts, >
Words.
I 204
In particular the material content of concepts is lost when the conceptual content is identified with the truth conditions. >
Truth conditions.
I 427/8
Definition Supervenience/Brandom: one vocabulary supervenes another if and only if there could be no two situations in which true assertions (i.e. facts) would differ expressably in the supervening vocabulary, while the true assertions do not differ expressably in the vocabulary that is being supervened more neutral: if it is clear what is defined in one language, then it is clear what is defined in the other. >
Supervenience.
I 958
Order/Twin Earth/TE/Brandom: it does not help to speak in concepts of what can be distinguished by the individuals, because what they can react depends on which reactions are considered to be different, and then the same problem occurs with regard to the vocabulary used Problem: specifying a vocabulary that satisfies two conditions:
1) The twins are indistinguishable in different environments because of their description in that vocabulary (physical language is not sufficient for that).
2) The sub-determination of the semantic properties of their states in this limited vocabulary must point at something interesting.
---
II 76
Material inference/Sellars/Brandom: from "a east of b" to "b west of a" also from flash to thunder, needs no logic.
II 79
Formally valid ones can be derived from good material inferences, but not vice versa Proof: if a subset of somehow privileged vocabulary is given, such an inference is correct if it is materially good and it cannot become a bad one if non-privileged vocabulary is replaced by privileged vocabulary. If one is only interested in logical form, one must be able to distinguish a part of the vocabulary as a especially logical beforehand. E.g. if one wants to explore theological inferences, one must investigate which replacement of non-theological vocabulary with non-theological preserves the material quality of the inference.
II 94
Definition "tonk"/Belnap: Rule 1): licenses the transition from p to p tonk q for any q.
Rule 2): licenses the transition from p tonk q to q. With that we have a "network map" for inferences: any conclusion is thus permitted.
PriorVsBelnap: Bankruptcy of all definitions in the style of Gentzen.
BelnapVsPrior: Solution: Restriction: no inferences with only old vocabulary that were not allowed previously,otherwise the old contents would be changed retrospectively.
