In his dissertation, MacFarlane presents some reasons for thinking that the view that Kant was the first to hold that what is distinctive of logic is that it is normative for thought as such and that logic’s abstraction from content is a consequence of this. To do this, he goes through some of the views of logic held by Kant’s predecessors. He wants to show that Kant agreed with his predecessors to some degree about logic so that his own views about logic would not be seen as merely changing the subject. Part of what made this exercise interesting for me was seeing what other views of logic were in play.

We start with the Kantian view that logic is normative for thoughts as such. Logic consists of a bunch of norms or rules for concept application. The judgment that all A are B says that in applying the concept A one ought to apply the concept B. Frege likewise thinks that logic is normative for thought. This drops out of the picture by the writing of the Tractatus. I’m fairly certain that there isn’t any normative status attributed to logic in the Tractatus. I’m not sure if, say, the Principia mentions norms at all.

Next, there is the Wolffian view of logic which holds, like Kant, that logic is normative for thought and also that logic can tell us substantive things about the world. It should, quoting MacFarlane, “be grounded in ontology and psychology.” One derives the rules of logic by examining psychology. The rules are also supposed to “be derived from the cognition of being in general, which is taken from ontology.” (Quoting Wolff.) Based on the examples of Kant’s criticisms of this view in MacFarlane’s dissertation, it seems like the adherents of it do not distinguish between rules applying to concepts and those applying to objects.

The Wolffians held that the form of thought was the same as the form of being, to use MacFarlane’s characterization. They thought that the relations among concepts were an accurate guide to the relations among beings. Possibly because of this they seemed to think that logic could be a guide to ontology. I’m not sure to what extent this meshes with contemporary views about logic, if at all. Alas, I’m not likely to read a bunch of Wolff in order to find out. This also, I expect, runs roughshod over distinctions that could be, and need to be, made about the phrase “guide to ontology”.

Third, there is the view held by Locke and Descartes: logic is a set of rules “which teaches us to direct our reason with a view to discovering the truths of which we are ignorant.”Interestingly, Descartes criticizes what he characterizes as Scholastic logic for being about mere forms of reasoning, unconcerned with the truth of the premises of the arguments. This view of logic holds that it is normative, not necessarily for thought as such, and must be grounded in empirical psychology. There could be thought that does not follow the norms of this logic; it just wouldn’t be particularly useful for generating knowledge. Logic is supposed to extend our knowledge in substantive ways. It is also supposed to help us avoid sophisms since we are not concerned just with mere forms.

It isn’t clear what the continuity is between these views of logic and the current one(s). A link might be obtainable with the Kantian one via Frege. Frege thought that the laws of logic were the laws of thought, but his laws of logic took a form more familiar to us than Kant’s. This seems like a family resemblance sort of transition since we lose the normativity in the process. Setting that aside, if something of the older Kantian view of logic is lingering in the modern conception of logic, then it seems like there is plenty of room for there to be logical notions, completely distinct from mathematical ones. Although, this notion of being completely distinct from needs clarification, possibly along the lines sketched by Colin for the notion of dependence in the comments a few posts ago.